Dædalus Coming up in Dædalus

Dædalus coming up in Dædalus:

On Water Christopher Field & Anna Michalak, Michael Witzel, Charles Vörösmarty, Michel Meybeck & Christopher L. Pastore, Terry L. Dædalus Anderson, John Briscoe, Richard G. Luthy & David L. Sedlak, Stephen R. Carpenter & Adena R. Rissman, Jerald Schnoor, Katherine Jacobs, and Journal of the American Academy of Arts & Sciences others Spring 2015

Food, Health & G. David Tilman, Walter C. Willett, Meir J. Stampfer & Jaquelyn L.

the Environment Jahn, Nathaniel D. Mueller & Seth Binder, Andrew Balmford, Rhys Spring 2015: Successful Aging of Societies Green & Ben Phalan, G. Philip Robertson, Brian G. Henning, and others Successful John W. Rowe Successful Aging of Societies 5 Aging of S. Jay Olshansky The Demographic Transformation The Internet David Clark, Yochai Benkler, Peter Kirstein, Deborah Estrin & Ari Societies of America 13 Juels, Archon Fung, Susan Landau, John Palfrey, and others Robert A. Hummer Hispanic Older Adult Health & Longevity in the United States: Current Patterns & What’s New Matthew S. Santirocco, Shadi Bartsch, Angelos Chaniotis, Walter & Mark D. Hayward Concerns for the Future 20 About the Old? Scheidel & Ian Morris, Phillip Mitsis, Roger Bagnall, Verity Platt, Emily Greenwood, Peter T. Struck, Greg Crane, Caroline Alexander, Frank F. Furstenberg, The Future of Intergenerational Relations Brooke Holmes, Kyle Harper, and others Caroline Sten Hartnett, in Aging Societies 31 Martin Kohli plus New Dilemmas in Ethics, Technology & War; Political & Julie M. Zissimopoulos Leadership &c Lisa F. Berkman, Labor-Force Participation, Policies & Practices Axel Boersch-Supan in an Aging America: Adaptation Essential & Mauricio Avendano for a Healthy & Resilient Population 41 Dawn C. Carr, Productivity & Engagement in an Aging Linda P. Fried America: The Role of Volunteerism 55 & John W. Rowe S. Jay Olshansky, Resetting Social Security 68 Dana P. Goldman & John W. Rowe David E. Bloom, Global Population Aging: Facts, Challenges, David Canning Solutions & Perspectives 80 & Alyssa Lubet Julie M. Zissimopoulos, Individual & Social Strategies to Mitigate Dana P. Goldman, the Risks & Expand Opportunities S. Jay Olshansky, of an Aging America 93 John Rother & John W. Rowe

U.S. $14; www.amacad.org Cherishing Knowledge · Shaping the Future

Inside front cover: Indian photographer G. B. Mukherji’s award- winning photograph “Tibetan Grandfather” portrays how tight intergenerational family bonds can facilitate communal security and cohesion. Although Tibetans living in exile in India face an uncertain future, their devotion to family serves as a source of strength and hope. © G. B. Mukherji/Generations United. John W. Rowe, Guest Editor Phyllis S. Bendell, Managing Editor and Director of Publications D Peter Walton, Assistant Editor Emma Goldhammer, Senior Editorial Assistant J

Committee on Studies and Publications John Mark Hansen and Jerrold Meinwald, Cochairs; Denis Donoghue, Gerald Early, Carol Gluck, Sibyl Golden, Linda Greenhouse, John Hildebrand, Jerome Kagan, Philip Khoury, Arthur Kleinman, Steven Marcus, Rose McDermott, Eric Sundquist, Jonathan F. Fanton (ex of½cio), Don M. Randel (ex of½cio), Diane P. Wood (ex of½cio)

Dædalus is designed by Alvin Eisenman. Dædalus Journal of the American Academy of Arts & Sciences

Design for the hedge maze is by Johan Vredeman de Vries, from Hortorum viridariorumque elegantes & multiplices formae: ad archi- tectonicae artis normam affabre delineatae (Cologne, 1615).

Dædalus was founded in 1955 and established as a quarterly in 1958. The journal’s namesake was renowned in ancient Greece as an inventor, scientist, and unriddler of riddles. Its emblem, a maze seen from above, symbol- izes the aspiration of its founders to “lift each of us above his cell in the lab- yrinth of learning in order that he may see the entire structure as if from above, where each separate part loses its comfortable separateness.” The American Academy of Arts & Sciences, like its journal, brings together distinguished individuals from every ½eld of human endeavor. It was char- tered in 1780 as a forum “to cultivate every art and science which may tend to advance the interest, honour, dignity, and happiness of a free, independent, and virtuous people.” Now in its third century, the Academy, with its nearly ½ve thousand elected members, continues to provide intellectual leadership to meet the critical challenges facing our world. Dædalus Spring 2015 Subscription rates: Electronic only for non- Issued as Volume 144, Number 2 member individuals–$47; institutions–$129. Canadians add 5% gst. Print and electronic for © 2015 by the American Academy nonmember individuals–$52; institutions– of Arts & Sciences $144. Canadians add 5% gst. Outside the United Editorial of½ces: Dædalus, American Academy of States and Canada add $23 for postage and han- Arts & Sciences, 136 Irving Street, Cambridge ma dling. Prices subject to change without notice. 02138. Phone: 617 576 5085. Fax: 617 576 5088. Institutional subscriptions are on a volume- Email: [email protected]. year basis. All other subscriptions begin with the next available issue. Library of Congress Catalog No. 12-30299. Single issues: $14 for individuals; $36 for insti- Dædalus publishes by invitation only and as- tutions. Outside the United States and Canada sumes no responsibility for unsolicited manu- add $6 per issue for postage and handling. scripts. The views expressed are those of the Prices subject to change without notice. author of each article, and not necessarily of the American Academy of Arts & Sciences. Claims for missing issues will be honored free issn e-issn of charge if made within three months of the Dædalus ( 0011-5266; 1548-6192) publication date of the issue. Claims may be is published quarterly (winter, spring, summer, submitted to [email protected]. Members of fall) by The mit Press, One Rogers Street, Cam- ma the American Academy please direct all ques- bridge 02142-1209, for the American Academy tions and claims to [email protected]. of Arts & Sciences. An electronic full-text version of Dædalus is available from The mit Press. Advertising and mailing-list inquiries may be Subscription and address changes should be ad - addressed to Marketing Department, mit Press dressed to mit Press Journals Customer Service, Journals, One Rogers Street, Cambridge ma One Rogers Street, Cambridge ma 02142-1209. 02142-1209. Phone:617253 2866. Fax: 617 253 1709. Phone: 617 253 2889; U.S./Canada 800 207 8354. Email: [email protected]. Fax: 617 577 1545. Email: [email protected]. To request permission to photocopy or repro- Printed in the United States of America by duce content from Dædalus, please complete the Cadmus Professional Communications, Science online request form at http://www.mitpress Press Division, 300 West Chestnut Street, journals.org/page/permissionsForm.jsp, or con- Ephrata pa 17522. tact the Permissions Manager at mit Press Jour- nals, One Rogers Street, Cambridge ma 02142- Newsstand distribution by Ingram Periodicals tn 1209. Fax: 617 253 1709. Email: journals-rights@ Inc., 18 Ingram Blvd., La Vergne 37086. mit.edu. Postmaster: Send address changes to Dædalus, Corporations and academic institutions with One Rogers Street, Cambridge ma 02142-1209. ma valid photocopying and/or digital licenses with Periodicals postage paid at Boston and at the Copyright Clearance Center (ccc) may additional mailing of½ces. reproduce content from Dædalus under the The typeface is Cycles, designed by Sumner terms of their license. Please go to www Stone at the Stone Type Foundry of Guinda ca. .copyright.com; ccc, 222 Rosewood Drive, Each size of Cycles has been sep arately designed Danvers ma 01923. in the tradition of metal types. Successful Aging of Societies

John W. Rowe

Abstract: As America ages, policy-makers’ preoccupations with the future costs of Medicare and Social Security grow. But neglected by this focus are critically important and broader societal issues such as inter - generational relations within society and the family, rising inequality and lack of opportunity, productivity in late life (work or volunteering), and human capital development (lifelong education and skills training). Equally important, there is almost no acknowledgment of the substantial bene½ts and potential of an aging society. The MacArthur Foundation Research Network on an Aging Society offers policy options to address these issues and enhance the transition to a cohesive, productive, secure, and equitable aging society. Such a society will not only function effectively at the societal level but will provide a context that facilitates the capacity of individuals to age successfully. This volume comprises a set of papers, many of which are authored by members of the MacArthur Network, focusing on various aspects of the opportunities and challenges facing the United States while it passes through its current demographic transformation. This essay provides a general overview of the strategy the Network has used to address the various components of this broad subject.

Policy-makers and pundits are increasingly preoc- cupied with the negative economic effects of population aging on public health and pension entitlements, including Medicare and Social Security. The enormous unfunded future obligations of these programs, especially Medicare, tend to crowd out JOHN W. ROWE, a Fellow of the all other considerations. While these entitlement American Academy since 2005, is programs surely require modi½cations to ensure Professor at the Columbia Univer- their sustainability and fairness, the current debate sity Mailman School of Public Health and Chair of the MacArthur neglects other critically important issues related to Foundation Re search Network on the aging of America: future intergenerational re - an Aging Society. He is the author lations and tensions; socioeconomic disparities and of Successful Aging (with Robert L. inequalities; changes in the structure and function Kahn, 1998) and was the Chair of of the family and its capacity to serve the traditional the Institute of Medicine of the Na- safety-net role; the impact of technology; and the tional Academies project the Future critical importance of adaptation of core societal Health Care Workforce for Older Americans, which authored the re- institutions, including education, work and retire- port Re tooling for an Aging America: ment, housing, transportation, and even the design Building the Health Care Workforce of the built environment (the supporting residential, (2008). recreational, commercial, and transportation infra-

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00325 5 Successful structure). Equally important, there is al - The influence of the baby boom on U.S. Aging of most no acknowledgment of the substan- population aging is not temporary. Con- Societies tial positive contributions and potential trary to what the popular myth suggests, productivity of an aging society. the passing of the baby boomers through Our goal is to develop and help imple- the age structure will not terminate pop- ment policies that assure our transition to ulation aging or return us to the age struc- a cohesive, productive, secure, and equi- ture of earlier periods of U.S. history. Rath- table aging society. Failure to reach this er, the demographic changes that have goal will leave us with a society rife with taken place over the last century are per- intergenerational tensions–characterized manent. The age structure of all current by enormous gaps between the haves and and future populations either have already the (increasingly less-educated) have-nots been transformed or are about to perma- in quality of life and opportunity–and nently shift, aggravated in part by the un- unable to provide needed goods and ser - usually large post–World War II birth vices for any of its members, especially a cohort, but driven primarily by the com- progressively older and more dependent bined effect of unprecedented increases in population. life expectancy and decreases in birth rates. Gloomy though this scenario is, it is The second widely accepted myth is that avoidable. We have time to put in place an aging society is de½ned by and is solely policies that will help strengthen the fu - concerned with its elders. This belief tends ture workforce, increase productive en- to pit generations against each other, over - gagement of older individuals, and enhance looking the critical fact that the proper the capacity of families to support elders. unit of analysis for policy-makers is not Many such policies may, at the same time, one speci½c age cohort but rather society lessen the burden on Social Security. as a whole. Policy-makers must consider the intergenerational effects of their poli- How did we get here? Given the advance cies and design solutions that bene½t all warning decades ago that an age wave was of society, not just any one interest group. coming, why has U.S. society been unable Whereas countries in Western Europe to prepare? Part of the failure to act lies aged ahead of the United States–reflecting with a set of archaic beliefs regarding the their post–World War II baby bust and true nature of societal aging. Stakehold- sustained reductions in total fertility be- ers failed to realistically assess challenges low the replacement rate–the U.S. baby and envision opportunities and squan- boom and higher fertility rate have com- dered the time available to formulate ap- bined to delay by a few decades the emer- propriate public policy. The denial con- gence of an aging society (de½ned here as tinues: a recent Pew Research Center sur- one with more individuals over age sixty- vey of global attitudes on aging shows that ½ve than are under age ½fteen). For in- less than 26 percent of Americans feel that stance, the United States will not meet Ger- an aging society is a “major issue”! Only many’s current population age distribu- Indonesia and Egypt ranked lower on the tion until 2030. And Germany’s age struc- survey.1 Contributing to this denial are ture has not caused ruin for its society or two pervasive and disabling myths about its economy. Thus, one would think that aging in the United States: the ½rst myth the experiences of the Western European concerns the impact of the baby boom; countries, which are like the United States the second assumes that an aging society in many ways, would provide a clear road is only concerned with the elderly. map for the policies the United States

6 Dædalus, the Journal ofthe American Academy of Arts & Sciences needs to adopt for a successful transition ment, childrearing, and leisure) across the John W. to a productive and equitable aging society. individual life span. Stakeholders need to Rowe But although the United States certainly detail the impact of socioeconomic, racial/ has much to learn from looking at the ethnic, and gender differences on life- experiences of older societies in Europe course trajectories and specify how they and even of Japan, differences across so - influence the effectiveness of various life - cieties, cultures, and policy strategies may style related interventions. limit the utility of these comparisons, 4) Consider bene½ts and risks. Analysis of thus requiring the development of a policy changes should consider both the uniquely American resolution to the is- possible bene½ts and risks to an aging so - sues presented by an aging society. In ciety and should develop a unifying strat- short, international comparisons can be egy that optimizes the balance between valuable, but we must be cautious in gen- the two. As societies attempt to deal with eralizing experiences from other cultures. the many challenges derived from demographic transition, too little attention is The MacArthur Network has developed paid to its potential upside: the longevity a set of closely related components that dividend. This includes the previously un- form the core of a theory of adaptation in imaginable capability of older individuals an aging society. Although there is substan- to participate productively in society either tial overlap between these components, through the workforce or through civic identifying each has value. To begin, a plan en gagement. Older people have much to of action must ½rst: offer, including accrued knowl edge, stabil- 1) Analyze society and its institutions. The ity, unique creative capacities for synthetic unit of analysis should be the society and problem solving, and increased ability to the adaptation of its core institutions (such manage conflicts and consider the perspec- as family, work and retirement, educa- tives of other age groups. As a society, the tion, media, religion, and civic affairs) and United States should harness the life-stage- should encompass a multigenerational and appropriate capabilities and goals of peo- intergenerational perspective, rather than ple of all ages, including older adults, to en- focus solely on individuals of any one age hance societal bene½ts and reduce social group (elders or youth). strati½cation. 2) Take a long-term view and consider struc- 5) Focus on human capital. Policy-makers tural lag. The primary focus should be on should focus on strategies that take ad- adjusting and adapting core institutions– vantage of all available talent in the popula- including education, work and retirement, tion, employ social norms based on ability health care, the design and function of rather than chronological age, and transi- housing and cities, and transportation– tion from an emphasis on investment early over the long term. It is important to keep in life to recognition that investments in mind gerontologist Matilda Riley’s con- across the full life span can pay dividends. cept of structural lag: the recognition that These payoffs will be individual, intergen- most societal institutions are resis tant to erational, and societal (with both cross- change and lag behind the shifting popu- over and spillover effects); and because lation of their members.2 they can be positive or negative, the out- 3) Adopt a life-course perspective. U.S. so - comes must be monitored. ciety needs to adopt a life-course perspective that urges redistribution of life’s ac- The MacArthur Network has developed tivities (such as education, work, retire- three strategies for policy analysis. First,

144 (2) Spring 2015 7 Successful it is critical to develop a toolbox of more widening gap between the haves and have- Aging of sensitive and predictive economic and so - nots and for the increased competition Societies cial indicators–including lifestyle dimen- over scarce resources being channeled into sions–that permit accurate assessment of entitlements to tear at the fabric of our the current conditions and likely future society and create a “war” between the trajectory of the population and society generations? along the principal policy dimensions of The MacArthur Network prefers to use interest. We need an alternative to the the term cohesion to describe the issues re- archaic old-age dependency ratio, which lated to intergenerational relations (or simply equates old age with dependency. tensions) because it focuses on age inte- Metrics that express the full array of gration rather than age segregation and bene½ts-to-costs relationships of a long- addresses intergenerational transfers, at - lived society, as well as alternatives for life- titudes, multigenerational strategies, and course trajectories, are also essential. This changes in family structure. Cohesion can toolbox can be used to model possible out- be viewed as the debate regarding the tra- comes of societal investment in factors that ditional social compact–which we prefer alter the impact of an aging population. over the more commonly used legalistic Second, in order to encourage the iden- “contract”–between the generations. ti½cation of effective solutions, researchers Substantial empirical evidence shows and policy-makers must present and ana- strong support by middle-aged and youn - lyze multiple policy options, rather than ger Americans for older Americans and advocate single proposals, and should tar - highlights social cohesion’s bene½ts; but, get multiple factors (such as the ½nancial, as many observers have noted, the future social, life-course evolution, behavioral, increase of entitlement costs may place and physical). Further, policy-makers substantial stress on this balance.3 De - should consider and employ both private pending on future economic and educa- and public involvement and federal and tional gaps, will future young-adult and local approaches. middle-aged Hispanics, for example, re - Finally, policy analysis must assess pol- flect the same support for elderly white icy impacts. The MacArthur Network sug- Americans? Further, what impact will fu- gests adopting a strategy similar to that ture immigration policies, whose intent used to assess the environmental impact may be to eliminate the shortfall of skilled of a planned development. Speci½cally, U.S. workers, have on these tensions? Net work members propose that all poli- 2) Family (evolution, supports, changing cies be evaluated for the effects they have roles). Families make up the front line of within each generation, as well as on the our adaptation to an aging society. For the interactions between generations (known family, the core question of the aging soci- as assessing intergenerational effects), in ety relates to the uncertainty regarding its order to be most effective. capacity to play its traditional role as safety In addition, the MacArthur Network has net and exhibit adaptive capacities to re - identi½ed six high-priority domains for spond to a variety of ½nancial, social, and policy analysis. They include: health-related needs. Factors threatening 1) Intergenerational relations. This general the family’s role include the emergence area requires understanding at both the of an array of family forms with different societal and individual family-unit lev els. capacities for support (such as a childless For society, the core question relates to family unit), increased longevity, geo- cohesion. What is the potential for the graphic dispersion, economic challenges,

8 Dædalus, the Journal ofthe American Academy of Arts & Sciences and likely future reductions in entitle- egies–may have a substantial positive John W. ments. effect on post-retirement engagement. Rowe Moreover, these changes are ampli½ed Such engagement is bene½cial not only by the growing diversity that results from for retirees but also for the general popu- increased strati½cation. The strength and lation. salience of intergenerational ties become Technology bridges the worksite to areas more prominent features in an aging so - of civic engagement and, depending on ciety, and the traditional life course is the type of technology and its ½t with the being altered in part because of increased abilities and needs of older individuals, can longevity. The transition to adulthood wind up either facilitating or inhibiting comes ½ve or more years later than it used their participation. Substantial opportu- to, placing parents of young adults in the nity exists for policy changes and techno- challenging position of helping support logical and other worksite modi½cations their parents or even their grandparents and educational interventions that will not while launching their own children toward only make retention of older workers more independence.4 Families with resources attractive to employers, but will also take can manage this balancing act relatively advantage of the many strengths older well, but a growing number of families workers offer. It is important for policy- will be overly burdened trying to contend makers to be aware of the “lump-of-labor” with these competing demands without fallacy and the growing body of empirical proven ways of managing the more com- evidence indicating that older individuals plex, intergenerational family systems. need not be moved out of the workforce Issues such as intrafamilial supports, hous- to make room for younger workers.6 In ing, ½nancial transfers, caregiving, and addition, policy should be informed by the new familial roles will also inform critical most recent ½ndings regarding trends in policy decisions surrounding the chang- disability in populations of elders and ing face of U.S. families. near-elders. Much of the most recent work 3) Productivity (work and retirement, func- suggests that the severe disability rates tional status and disability, technology, roles of (as measured by activities of daily living older individuals in society). The future roles and instrumental activities of daily living of older individuals in society will have a scales) are now stable in older individuals, dramatic impact on the likelihood that having halted their decades-long decline; the United States will be productive, co - and that, for unknown reasons, functional hesive, and equitable. This set of issues can mobility impairments may be rising in in- be conveniently divided between work and dividuals aged ½fty to sixty-½ve.7 It will retirement mat ters and civic engagement be important for policy-makers to under- matters, although they are closely inter- stand the likely influence of these trends related. The likelihood of a retiree volun- on the adequacy of the future U.S. labor teering is very much influenced by wheth- force, as well as on the future demand for er that person volunteered while still in personal care services. the workforce.5 Thus, approaches to en - 4) Human capital development (such as life - couraging people to volunteer while still long education and skills training). Some of in the workforce–via modi½cations in the same societal forces that led to longer time and place of work, provision of op - lives have also shortened the half-life of portunities for en gaging in what individ - knowledge in science and technology. How uals consider meaningful activities, and can human capital be expanded at different development of paid volunteerism strat - points along the life course? Can the mis-

144 (2) Spring 2015 9 Successful alignment between education and work tainable and clearly articulated policies Aging of that is aggravated by increasing longevity that deal humanely with care at the end of Societies be improved through a closer relationship life. between educational institutions and the 6) Relevance to successful aging of individuals. workplace? Over the past ½fteen years, successful aging Stakeholders need to understand and has been a major theme of gerontological employ the most effective approaches to research. Much of the work in the ½eld has keep young individuals in school and to been stimulated by the model of success- provide a coherent approach to lifelong ful aging proposed by the MacArthur Net- learning that gives individuals the skills work on Successful Aging, which is fo- and attitudes they need to continue to cused primarily at the level of the individ- productively evolve within overall societal ual.8 It is self-evident that the changes that and work environments. Although return- occur at the societal level in response to ing to school–now common among youn- the demographic transformation may have ger adults–is still relatively rare among major positive or negative effects on the individuals over forty, providing access capacity of individuals to age successfully. to educational institutions for the near- While many of the issues and policy op - old and old is no less critical than keeping tions discussed in this volume are relevant younger people in school. Education must to individuals, our primary current focus be rede½ned as a lifelong experience. is at the level of society. The interaction 5) Health and health care. Although it between societal change and the status of might seem that the ongoing national de- aging individuals represents fertile terri- bate about health care reform may have tory for future research. exhausted this topic, the Network believes that some important and often neglected These major themes and recommenda- areas of the discussion are directly related tions are explored in depth in the essays to the demographic transformation. These found in this issue of Dædalus. Among the include the development of a more geri- essays are S. Jay Olshansky’s “The Demo- atrically sophisticated health care system graphic Transformation of America,” in which most providers (physicians, nurs- which looks toward the changing face of es, dentists, social workers, psychologists, aging and life expectancy in America. pharmacists, and others) are competent Robert Hummer and Mark Hayward’s in diagnosing and treating medical dis- essay “Hispanic Older Adult Health & Lon- eases and syndromes that are common in gevity in the United States: Current Pat- old age, as well as a strong reliance on new terns & Concerns for the Future” ex plores interdisciplinary models of care that are the “Hispanic paradox”–that ½rst-gener - more effective in managing the health care ation Hispanic immigrants have a greater problems of frail older individuals with life expectancy than both nonimmigrant multiple impairments. In addition, a reori- Americans and residents of their native entation to a life-course preventive health countries–in addition to troubling health model is needed to strengthen education and well-being warning signs for the fu- about healthy lifestyles and intervention ture Hispanic population. Frank Fursten- implementation in at-risk groups so that berg, Caroline Hartnett, Martin Kohli, and future older individuals will enter the Julie Zissimopoulos have written “The Fu- Medicare program healthier and at higher ture of Intergenerational Re lations in levels of functioning than their predeces- Aging Societies,” which examines the fam- sors. Finally, the United States needs sus- ily’s capacity to respond to the growing

10 Dædalus, the Journal ofthe American Academy of Arts & Sciences challenges and demands for support of a In their essay “Global Population Aging: John W. rapidly aging America; while Lisa Berk- Facts, Challenges, Solutions & Perspec- Rowe man, Axel Boersch-Supan, and Mauricio tives,” David Bloom, David Canning, and Avendano point toward how adaptation Alyssa Lubet provide an overview of global of our expectations of the elderly can lead population aging and its contributing fac- to a more productive and resilient society tors; outline some of the major challenges in “Labor-Force Participation, Pol icies & associated with widespread population Practices in an Aging America: Adaptation aging; and describe current and possible Essential for a Healthy & Resilient Popu- future responses to these challenges. Fi- lation.” nally, Julie Zissimopoulos, Dana Goldman, In our essay “Productivity & Engage- S. Jay Olshansky, John Rother, and I con- ment in an Aging America: The Role of clude the issue with “Individual & Social Volunteerism,” Dawn Carr, Linda Fried, Strategies to Mitigate the Risks & Ex pand and I propose that the impact of volun- Opportunities of an Aging America.” This teerism in an aging population be recog- essay discusses the major risks associated nized and invested into, and that programs with aging at both the level of the individ- harness the social capital of older adults to ual and the level of society, and presents improve the well-being of the elderly and courses of action for policy-makers in edu- address critical needs of society as a whole. cation, work and retirement, ½nancial se- And S. Jay Olshansky, Dana Goldman, and curity, health care, and social cohesion to I contributed the essay “Resetting Social promote the bene½ts and reduce the risks Security,” which considers the critical ½ - of longer life. Taken together, these policy nan cial safety net of social security and options provide a broad blueprint for suc - what impact might result from further cessful societal adaptation to the aging of changes to its age of eligibility require- America. ments.

endnotes 1 Pew Research Center, Attitudes about Aging: A Global Perspective (Washington, D.C.: Pew Re - search Center, 2014). 2 Matilda Riley and John Riley, “Structural Lag: Past and Future,” in Age and Structural Lag: Society’s Failure to Provide Meaningful Opportunities in Work, Family, and Leisure, ed. Matilda Riley, Robert L. Kahn, and Ann Foner (New York: Wiley-Interscience, 1994). 3 James Schulz and Robert Binstock, Aging Nation: The Economics and Politics of Growing Older in America (Baltimore: Johns Hopkins University Press, 2008). 4 Gordon Berlin, Frank F. Furstenberg, Jr., and Mary Waters, “The Transition to Adulthood,” The Future of Children 20 (1) (Spring 2010): 1–18. 5 Barbara A. Butrica, Richard W. Johnson, and Sheila R. Zedlewski, “Volunteer Dynamics of Older Americans,” The Journals of Gerontology: Psychological Sciences & Social Sciences 64 (5) (February 2009): 644–655. 6 Axel Boersch-Supan, Reduction of Working Time: Does it Decrease Unemployment? mea Dis- cussion Paper No. 2003 (Mannheim, Germany: University of Mannheim, Mannheim Re - search Institute for the Economics of Aging, 2002); and John Gruber and David Wise, eds., Social Security Programs and Retirement Programs Around the World: The Relationship to Youth Employment (Chicago: University of Chicago Press, 2009).

144 (2) Spring 2015 11 Successful 7 National Research Council, “Health and Disability in the Working-Age and Elderly Popula- Aging of tions,” in Aging and the Macroeconomy: Long-Term Implications of an Older Population (Wash- Societies ington, D.C.: National Academies Press, 2012). 8 John W. Rowe and Robert L. Kahn, “Human Aging: Usual and Successful,” Science 237 (4811) (1987): 143–149; John W. Rowe and Robert L. Kahn, “Successful Aging,” The Gerontologist 37 (4) (1997): 433–440; and John Wallis Rowe and Robert L. Kahn, Successful Aging (New York: Pantheon Books, 1998).

12 Dædalus, the Journal ofthe American Academy of Arts & Sciences The Demographic Transformation of America

S. Jay Olshansky

Abstract: The face of aging in America is about to change. Within the next thirty years, the U.S. population will experience a permanent change in its age structure, and there is reason to believe that cohorts reaching older ages in the future will be far different from those reaching older ages today. In this essay, I explain why life expectancy in the United States is likely to diverge from that experienced by the rest of the developed world; describe recent trends in healthy life expectancy; and examine how the age structure of the United States will by mid-century be different from that found today.

The face of aging in America is undergoing a profound transformation. Within the next thirty years, the U.S. population will accompany the rest of the developed world in experiencing a permanent change in its age structure. The United States will not only be much older in the coming decades, con- fronting a suite of resulting challenges and opportunities, but cohorts that reach older ages in the fu - ture are likely to be far different from those reaching older ages today. The reason? Older cohorts in the future will have been born into and lived through an entirely different set of environmental and medical/ health conditions relative to their counterparts born S. JAY OLSHANSKY is Professor in the early twentieth century. In this essay, I begin of Epidemiology at the School of by explaining why life expectancy in the United Public Health, Division of Epidemi - States is likely to diverge from that experienced by ology and Biostatistics at the Uni- the rest of the developed world (a phenomenon that versity of Illinois at Chicago. He has has already begun); describe recent trends in healthy published articles in such journals life expectancy; and examine how the age structure as The New England Journal of Medi- cine, JAMA: The Journal of the Amer- of the United States by mid-century will be different ican Medical Association, Science, The from that found today. Weaved into the discussion Sci entist, Sci enti½c American, and of each topic is the issue of how disparities are likely Health Affairs. to influence trends in longevity and health.

13 The The question of how high life expectan- medical technology yielded reductions in Demo - cy in the United States can rise has been case fatality rates for people with cardio- graphic Transfor - the subject of intense debate for decades. vascular diseases, cancer, and diabetes. mation of Because forecasts of longevity influence How ever, the resulting gain in life expec - America efforts to ensure the ½nancial integrity of tancy at birth as a product of declining age entitlement programs such as Social middle-age mortality was much smaller Security and Medi care, the answer to this than that observed in the early twentieth question has im portant public policy century because of entropy in the life table: implications. In order to understand declining death rates at middle ages and what the future might bring, we must above yielded progressively smaller gains ½rst place our current longevity within in life expectancy since the total person- historical context. years-of-life added to the life table is much The modern rise in life expectancy is one smaller than what occurs when saving of humanity’s crowning achievements. the young.2 Historical trends suggest that after more Long-lived populations such as the than two hundred thousand years of slow United States’ are now in a position where but steady increases, a new chapter in the the only way to signi½cantly increase life book of human longevity began in the mid- expectancy in the future is to generate dra- dle of the nineteenth century with a quan- matic reductions in death rates at the old- tum leap in average duration of life. As est ages and simultaneously push the en - modern populations learned how to insu- velope of survival into the outer regions of late themselves from the hazards of the the lifespan (ages above one hundred twen - outside world, the external forces of mor- ty), where only a handful of people have tality (infectious diseases, predation, and ever lived. That is, large increases in life ex - accidents) that precluded survival beyond pectancy at birth in the future require not the ½rst few years of life were for most only large declines in death rates for older people largely relaxed.1 The rapid increase people on par with what was ob served in in life expectancy was initially due to ad- the past for young and middle-aged popu- vances in public health (such as refrigera- lations; it also requires that most people tion, sewage disposal, clean water, and in - either routinely live past the age of one door living and working environments) hun dred ten, or that a signi½cant segment that saved the lives of the young. When the of the population begin surviving well past lives of young people are extended, life the age of one hundred thirty. Is this likely expectancy rises rapidly since a large num- to happen? Unfortunately, the an swer is ber of person-years-of-life are added to the no, and there are three reasons why. total population (a phenomenon that can First, it is now acknowledged that the bi - only occur once). Once reductions in early- ological processes of aging represent the age mortality are achieved, future gains in most important risk factor for fatal dis- life expectancy must then be a product of eases expressed at older ages.3 Since it is reductions in death rates in other (middle not currently possible to signi½cantly alter and older) regions of the lifespan. the processes of aging, there is no reason The latter part of the twentieth century to suspect that dramatic declines in death followed this model exactly: death rates rates among the extreme elderly are plau- at middle ages began to decline as some sible. To the contrary, entropy in the life behavioral risk factors for the U.S. popu- table is likely to continue to erode gains lation improved (such as reductions in in life expectancy in the future. This does smoking prevalence) and as advances in not mean declines in death rates at older

14 Dædalus, the Journal ofthe American Academy of Arts & Sciences ages cannot be achieved–they can–rather, The bottom line regarding the future of S. Jay it just means that the resulting gains in life longevity in the United States is that there Olshansky expectancy will be small. is no reason to expect that survival can be Second, the age distribution of death in routinely pushed beyond the age of one long-lived populations like that of the hundred thirty; there are no med ical Unit ed States has indeed shifted to later breakthroughs on the horizon that offer ages, but this shift has been characterized the prospect of radical life extension; and by a compression of death into a fairly nar- there is compelling evidence to suggest row region between the ages of sixty-½ve that population subgroups within the and ninety. There is no evidence that the United States are simultaneously moving prospects for surviving past the age of one in opposite directions with regard to future hundred ten are improving; there is no longevity. Overall, there is reason to ex pect reason to expect people will routinely live that life expectancy in the United States beyond the age of one hundred thirty, an may rise marginally in the coming decades, age to which no human in history has been and that what is far more important from documented to live; nor is there reason to a public health perspective is how healthy believe that in a genetically heterogeneous the population will be in the future. population, everyone has the potential to live as long as the longest-lived member Life expectancy at birth and at older ages of the population. The implications of has been rising in the United States at a these observations are straightforward: fairly steady pace for most of the last cen- mortality compression is the most likely tury; but of equal interest to researchers scenario going forward, and this, in turn, are the trends in how healthy U.S. popu- must be accompanied by a decelerating lations are along the way. The measure of increase in life expectancy. life expectancy is often taken as a barom- Finally, there are vast differences in eter of a population’s health, yet, in fact, it longevity prospects among existing birth is more appropriately de½ned as little more cohorts in the United States–a phenom- than a measure of death. Life expectancy enon well documented in the scienti½c lit- tells us nothing at all about how healthy erature.4 As detailed in the sections that people are when alive. The proportion of follow, some subgroups of the U.S. popu- total life expectancy lived in a state of good lation today are facing rather bleak health health free from frailty and disability is and longevity prospects for the future. If known as healthspan and is mea sured by these health and longevity trends play out a metric referred to as healthy life expec- as suspected, some people will live longer tancy (hle). Because the unique data re - and healthier lives relative to populations quired to calculate it has only recently be- alive today, but others may very well ex - come available, measures of hle have only perience a decline in life expectancy on been calculated for developed nations par with observed reductions in life expec- since the early 1970s. tancy among the least educated white men Trends in hle for the United States in - and women in the United States.5 The dicate that a complex pattern has emerged His panic paradox, discussed by Robert consistent with the complexity of how life Hummer and Mark Hayward in their con - expectancy has changed in recent decades. tribution to this volume, suggests that this In 1970, hle at birth in the United States growing segment of the U.S. population was 67 for men and 74.6 for women, ris- could place an additional dampening effect ing to 71.8 and 78.8 for men and women, on the historic rise in life expectancy.6 respectively, by 1990.7 The rate of improve-

144 (2) Spring 2015 15 The ment in hle accelerated during this time of functionally independent older cohorts. Demo - frame, and the same trend toward increases In fact, there is evidence to indicate that a graphic hle Transfor - in and accelerated improvements also surprisingly large percentage of the pop- mation of occurred among the population aged sixty- ulation aged sixty-½ve and above in the America ½ve and above. However, when hle is con- United States today is physically and men- sidered within the frame of reference of tally operating at a level of ef½ciency that secular trends in total life expectancy, the is not far different from people who are 1970s were characterized by a slight expan- decades younger.12 sion of morbidity (when the rise in life ex - The challenging news is that disability pectancy outpaces the rise in hle), while rates have leveled off among people reach- the 1980s were characterized by a slight ing older ages during the last decade,13 and com pression of morbidity (when the rise recently published cohort studies sug gest in hle occurs faster than the rise in life ex - that younger cohorts moving through the pectancy).8 age structure are less healthy than their re- Given a history of known disparities in cent predecessors.14 Particularly notable is the life expectancies of American sub- the rise in pathology among children who groups,9 it should be no surprise that hle are both obese and who have diabetes–a is also inequitably distributed. Based on disturbing trend that does not bode well trends in hle at age thirty by race, sex, and for this generation. It implies that as con- level of completed education from 1970 to temporary younger and middle-aged gen- 1990, the gap in hle between population erations move into older regions of the life - subgroups is large. For example, in 1970 the span in the coming decades, unless these hle at age thirty for black men with less public health issues are ameliorated, we than a high school degree was 23.6 years, may witness declines in both hle and pos - while the hle at age thirty for white wom - sibly even life expectancy for the entire en with more than sixteen years of educa- U.S. population.15 Changing de mographic tion was 40.1 years.10 Interestingly, the gap conditions–discussed below–will influ- in hle between these two population sub - ence these trends. groups declined from 1970 to 1990; that is, hle increased by 0.2 years for the most The U.S. population has undergone a dra - educated white women while the least edu- matic demographic shift since the begin- cated black men experienced an increase ning of the twentieth century, when our in hle of 4.8 years. Nevertheless, differ- age structure was in the shape of a pyra- ences in total life expectancy and hle re - mid: few people reached older ages, con- main extremely large among population trasted by a comparatively large number subgroups demarcated by level of complet- of young people. By way of illustration, in ed education, essentially placing today’s 1900 the proportion of the total U.S. pop- disadvantaged populations in the middle ulation aged sixty-½ve and above was 4.1 part of the twentieth century in terms of percent; but this has risen to 14 percent their health and longevity prospects. today, and will rise to over 20 percent by The good news is that overall, the health mid-century.16 When rapid increases in status of older Americans has improved longevity combined with declining fertility during the last half century.11 The age- in the latter part of the twentieth century, speci½c incidence of functional impair- the U.S. age structure began shifting to a ments de½ned by activities of daily living more rectilinear form. By 2050, the age (adls) improved from 1980 to 2000, lead- structure of the United States and all other ing to notable increases in the prevalence developed nations will be in the shape of

16 Dædalus, the Journal ofthe American Academy of Arts & Sciences a square, with at least as many people population that is Hispanic will rise from S. Jay alive at older ages as there are at younger 16 percent today to 28 percent by 2050. Olshansky ages. This new shape to the U.S. age struc - More important, Hispanics now represent ture is likely to be a permanent feature of only 7 percent of the population aged sixty- our population for the foreseeable future. ½ve and above, but this will rise to 18 per- However, beneath the surface of a visi- cent by 2050. Neither of these projections bly shifting age structure are forthcoming would ordinarily be all that notable, ex - changes to our demographics that will alter cept for the fact that Hispanics represent the course of U.S. health and longevity by perhaps one of the more interesting anom- the mid-twenty-½rst century. Three major alies in U.S. demographics. events are now unfolding: First, there is Hispanics currently have the highest life evidence to suggest that subgroups of the expectancy among the main population U.S. population are experiencing signi½ - sub groups in the United States today. Hum - cantly different longevity and health tra- mer and Hayward demonstrate that this is jectories.17 While the least educated among due to the fact that the Hispanic population us compose a slowly shrinking seg ment of in the United States is currently dominated the population, being less educated today by ½rst-generation immigrants known to is far more lethal now than it was just two possess healthier lifestyles than either the decades ago. This trend will not have a pro - citizens of their country of origin or the found influence on national vital statistics gen eral U.S. population. This has led to because the proportion of the total popu- what is commonly known as the His panic lation that falls into this category is rela- paradox: the unexpected observation that tively small; but it will be a health chal- Hispanic immigrants currently live longer lenge nonetheless. than the total resident population.18 A second factor that will influence U.S. What makes the Hispanic impact on U.S. age structure in the twenty-½rst century demographics even more interesting is the is the advances in public health and the likelihood that the health and longevity of biomedical sciences that are likely to yield this subgroup is expected to worsen in the improvements in health and longevity. In- coming decades. Recent evidence indicates cluded among them are continued efforts that second- and third-generation Hispan- to reduce smoking prevalence; greater suc- ics are experiencing notable declines in cess in the treatment of complications as- health due to the acquisition of increasing- sociated with obesity; traction beginning ly harmful behavioral risk factors such as in the battle against the rise of childhood smoking and obesity.19 Thus, if Hispan- obesity; and anticipated ad vanc es in aging ics are about to dramatically increase their science that could yield a notable extension proportion of U.S. demograph ics, and their of healthy life by mid-century. future health and longevity trajectory is Finally, one of the more interesting de - spiraling downward, then there is reason velopments in shifting U.S. demographics to believe that this will have a notable neg- is the anticipated dramatic increase in the ative impact on national vital statistics Hispanic population and the unique im - such as life and health expectancy. Such fu - pact it will have on national health and ture trends would be invisible to currently longevity over the next few decades. This popular forecasting models that rely exclu- development is discussed in detail in sively on historical trends. Hummer and Hayward’s essay in this volume, but for now it is important to recog- The United States–along with the rest nize that the proportion of the total U.S. of the world–is headed down an inevi-

144 (2) Spring 2015 17 The table one-way path toward population els in use today, while others may experi- Demo - aging. However, there are elements to the ence signi½cant challenges. It is distinctly graphic Transfor - U.S. demographic transformation to an possible that two Americas will emerge: mation of aged society that set us apart from other one characterized by privilege as de½ned America developed nations. Unlike France and by higher levels of education, income, and Japan, where life expectancy at older ages every other bene½t packaged with higher has risen rapidly in recent decades, the socioeconomic status; and another char- United States has been characterized by acterized by lack of education, low income, stagnating or slowly rising life expectancy poverty, and the lifelong challenges im- at birth and older ages. The United States posed by a lifetime of lower socioeconom- is a far more heterogeneous population ic status. Our shifting demographics will relative to most other nations, and the dif- have a powerful influence on these trends, ferences among us have been accentuated and chief among them will be the unique over time by often radically different liv- influence of a rising Hispanic population. ing conditions for subgroups of the popu- The prospects for an aging America are lation. distinctive and challenging. Exactly how The overall trend toward population this phenomenon plays out in the coming aging in the United States will take on decades is unclear since so many different unique characteristics relative to the rest factors can and will influence health and of the developed world. Although life ex - quality of life going forward. What is pectancy at birth and at older ages is ex - known with certainty is that the U.S. age pected to rise in this century, there will be struc ture will change dramatically in this competing events that will influence the century, and there is reason to believe that future course of both longevity and health. subgroups of the population will experi- Some will live longer and healthier lives ence vastly different health and longevity than anticipated by most forecasting mod- prospects.

endnotes 1 S. Jay Olshansky, Bruce A. Carnes, and Christine K. Cassel, “The Aging of the Human Species,” Scienti½c American 268 (4) (1993): 46–52. 2 Nathan Key½tz, Applied Mathematical Demography, 2nd ed. (New York: Springer, 1985). 3 Denham Harman, “The Aging Process: Major Risk Factor for Disease and Death,” Proceedings of the National Academy of Sciences 88 (12) (1991): 5360–5363. 4 S. Jay Olshansky, Toni Antonucci, Lisa Berkman, Robert H. Binstock, Axel Boersch-Supan, John T. Cacioppo, Bruce A. Carnes, Laura L. Carstensen, Linda P. Fried, Dana P. Goldman, James Jackson, Martin Kohli, John Rother, Yuhui Zheng, and John Rowe, “Differences in Life Expectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up,” Health Affairs 31 (8) (2012): 1803–1813. 5 Ibid. 6 See Robert A. Hummer and Mark D. Hayward’s essay in this volume, “Hispanic Older Adult Health & Longevity in the United States: Current Patterns & Concerns for the Future.” 7 Eileen M. Crimmins, Yasuhiko Saito, and Dominique Ingegneri, “Trends in Disability-Free Life Expectancy in the United States, 1970–1990,” Population and Development Review 23 (3) (1997): 555–572. 8 Ibid.

18 Dædalus, the Journal ofthe American Academy of Arts & Sciences 9 Evelyn M. Kitagawa and Philip M. Hauser, Differential Mortality in the United States: A Study in S. Jay Socioeconomic Epidemiology (Cambridge, Mass.: Harvard University Press, 1973). Olshansky 10 Eileen M. Crimmins and Yasuhiko Saito, “Trends in Healthy Life Expectancy in the United States, 1970–1990: Gender, Racial, and Educational Differences,” Social Science & Medicine 52 (11) (2001): 1629–1641. 11 Vicki A. Freedman, Brenda C. Spillman, Patti M. Andreski, Jennifer C. Cornman, Eileen M. Crimmins, Ellen Kramarow, James Lubitz, Linda G. Martin, Sharon S. Merkin, Robert F. Schoeni, Teresa E. Seeman, and Timothy A. Waidmann, “Trends in Late-Life Activity Limi - tations in the United States: An Update from Five National Surveys,” Demography 50 (2) (2013): 661–671. 12 David J. Lowsky, S. Jay Olshansky, Jay Bhattacharya, and Dana P. Goldman, “Heterogeneity in Healthy Aging,” The Journals of Gerontology: Biological Sciences & Medical Sciences 69 (6) (2013): 640–649. 13 Freedman et al., “Trends in Late-Life Activity Limitations in the United States.” 14 Eric N. Reither, S. Jay Olshansky, and Yang Yang, “New Forecasting Methodology Indicates More Disease and Earlier Mortality Ahead for Today’s Younger Americans,” Health Affairs 30 (8) (2011): 1562–1568; and Teresa E. Seeman, Sharon S. Merkin, Eileen M. Crimmins, and Arun S. Karlamangla, “Disability Trends Among Older Americans: National Health and Nutrition Examination Surveys, 1988–1994 and 1999–2004,” American Journal of Public Health 100 (1) (2010): 100–107. 15 S. Jay Olshansky, Douglas J. Passaro, Ronald C. Hershow, Jennifer Layden, Bruce A. Carnes, Jacob Brody, Leonard Hayflick, Robert N. Butler, David B. Allison, and David S. Ludwig, “A Potential Decline in Life Expectancy in the United States in the 21st Century,” The New England Journal of Medicine 352 (11) (2005): 1138–1145. 16 U.S. Department of Health and Human Services Administration for Community Living, “Ad - min istration on Aging,” http://www.aoa.gov/Aging_Statistics/future_growth/future_growth .aspx (accessed September 20, 2014). 17 Olshansky et al., “Differences in Life Expectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up.” 18 Kyriakos S. Markides, Raphael Samper-Ternent, and Soham Al Snih, “Aging and Health in Mexican Americans: Selected Findings from the Hispanic epese,” in Race and Social Problems, ed. Ralph Bands and Larry E. Davis (New York: Springer, 2015), 171–186. 19 Mark D. Hayward, Robert A. Hummer, Chi-Tsun Chiu, César González-González, and Rebeca Wong, “Does the Hispanic Paradox in U.S. Adult Mortality Extend to Disability?” Population Research and Policy Review 33 (1) (2014): 81–96.

144 (2) Spring 2015 19 Hispanic Older Adult Health & Longevity p in the United States: Current Patterns & v H Concerns for the Future w

Robert A. Hummer & Mark D. Hayward t ½ g p Abstract: The Hispanic population aged sixty-½ve and over–the most socioeconomically disadvantaged subset of America’s elderly–is projected to quintuple between 2012 and 2050. While current longevity p patterns for Hispanics relative to whites are favorable, old-age functioning and disability patterns for p Hispanics are unfavorable and have serious implications for caregivers; families; and local, state, and 2 federal governments. Troubling signs for the future Hispanic population (which are shared to varying degrees with other vulnerable groups) include the unresolved legal status of unauthorized immigrants, continued low levels of insurance coverage even after health care reform, some unfavorable trends in health behaviors, and continued disadvantages in educational attainment and income relative to whites. We urge policy-makers to deal with these potentially problematic health and well-being issues. Not doing p so could have detrimental consequences for the future of the Hispanic population as well as other at-risk h groups and, by extension, the U.S. elderly population as a whole.

l l emographic data make clear that Hispanics will D B play an increasingly prominent role in the overall t health pro½le of U.S. society as the twenty-½rst cen - t tury progresses. By the year 2000, Hispanics had be- t come the country’s largest minority group at just ROBERT A HUMMER b . is Centennial over 35 million people, or one-eighth of the total pop - H Commission Professor in the Lib- ula tion. Between 2000 and 2010, the Hispanic pop- eral Arts in the Department of So - u ciology and Faculty Research Asso- ulation grew an additional 43 percent to over 50 mil- ciate in the Population Research lion people and increased its share of the population Cen ter at the University of Texas at to 16.3 percent.1 Furthermore, the U.S. Census Bu- Austin. reau projects that the Hispanic population will grow t to 112 million by 2050 and account for 28 percent of MARK D. HAYWARD is Professor f the total population.2 Should the reality of future of Sociology, Centennial Commis- w pop ulation changes come anywhere close to Census sion Professor in the Liberal Arts, l and Faculty Research Associate in Bu reau projections–and at present we see no reason v the Population Research Center at why it will not–it is straightforward to see that the T the University of Texas at Austin. future health patterns of U.S. society will increasingly reflect those of the Hispanic population. (For a de - (*See endnotes for complete contributor t biographies.) tailed discussion of demographic changes in the His - i © 2015 by the American Academy of Arts & Sciences t doi:10.1162/DAED_a_00327 20 panic population, see S. Jay Olshansky’s es- terns a challenging endeavor. Following Robert A. say on demographic transforma tion in this this overview, we discuss four key issues Hummer & Mark D. volume.) –undocumented immigrant status, health Hayward In addition to growing precipitously, the insurance cover age, trends in important Hispanic population is rapidly aging and health behaviors, and continued socio - will make up a progressively larger share eco nomic status disad vantages–that will of the older population in the coming de- likely have im portant impacts on the fu - cades. Indeed, the Census Bureau projects ture health and longevity patterns of the that the population of Hispanics over sixty- rapidly aging Hispanic pop ulation. We ½ve will quintuple between 2012 and 2050, con clude by urg ing a forward-thinking growing from 3.1 million to 15.4 mil lion. In policy agenda that will have the greatest percentage terms, while Hispanics now chance of enhancing the health and lon- com pose just over 7 percent of the na tion’s gevity pro½le of America’s largest minority population over sixty-½ve, that ½gure is group in the decades ahead. projected to increase to over 18 percent in 2050.3 Despite having a much higher level of Rapid Hispanic population growth and poverty and substantially lower levels of aging mean that it is important to under- educational attainment and health insur- stand current health and longevity patterns ance coverage than whites, Hispanics cur - among Hispanics and consider how such rently live longer lives, on average, than patterns may change given the so cial, be - their more socioeconomically advantaged havioral, and policy contexts of the United counterparts. The combination of greater States. Simply put, a more complete under - Hispanic longevity in the context of lower standing of current Hispanic health and socioeconomic status has long been con- longevity patterns will un doubtedly shed sidered an epidemiologic paradox (often light on what future patterns may look like. called the “Hispanic paradox”).4 While the But at the same time, we cannot assume quality of Hispanic mortality data used to that the Hispanic health and longevity pat- demonstrate this epidemiologic paradox terns of tomorrow will necessarily reflect has been debated, recent very high-quality those of today; indeed, some of the social, studies using different data sets and meth- behavioral, and policy contexts in which odologies have convincingly documented Hispanic health and longevity patterns are Hispanics’ greater longevity.5 Table 1, for unfolding are un fortunately less than fav - example, provides estimates of life expec- orable. tancy at age sixty-½ve for Hispanics and This essay ½rst provides an overview of whites from two national data sources. The current longevity and health patterns for table illustrates that Hispanic life expec - the older Hispanic population. We make tancy at age sixty-½ve is about two years frequent comparisons with non-Hispanic lon ger than that of whites, with only mi - whites (hereafter, “whites”), the nation’s nor variations across data sources and by larg est and most socioeconomically ad - gender. vantaged demographic subpopulation. The Hispanic longevity advantage over This description is not straightforward, be- whites varies by both nativity and national cause Hispanics are heterogeneous in origin. While U.S.-born Hispanics have no their nativity and national-origin compo- appreciable longevity advantage relative sition. More over, complex data-quality to whites, foreign-born Hispanics exhibit is sues have made the accurate documenta - a substantial advantage. For example, es- tion of Hispanic longevity and health pat- timated life expectancy at age sixty-½ve for

144 (2) Spring 2015 21 Hispanic Table 1 Older Adult Estimates of Remaining Life Expectancy at Age Sixty-Five for Hispanics and Non-Hispanic Whites Health & from Two National Data Sources Longevity in the Women Men United States NVSS NHIS-LMF NVSS NHIS-LMF Hispanics 21.7 21.9 19.0 18.8 Non-Hispanic Whites 19.7 20.3 17.1 16.8

“nvss” represents National Vital Statistics System, 2006 data; “nhis-lmf” represents National Health Interview Survey-Longitudinal Mortality Follow-Up, 1989–2006 data. Sources: Elizabeth Arias, “United States Life Tables by Hispanic Origin,” Vital and Health Statistics Series 2, No. 152 (Hyattsville, Md.: National Center for Health Statistics, 2010), 1–35; and Joseph T. Lariscy, Robert A. Hummer, and Mark D. Hayward, “Hispanic Older Adult Mortality in the United States: New Estimates and An Assessment of Factors Shaping the Hispanic Paradox,” Demography 52 (1) (2015).

foreign-born Hispanic women is three smoke) have recently been found to be an years longer than for white women.6 The important explanation for the low mortali- most important reasons for the exception- ty rates exhibited by older Mexican im - ally favorable longevity patterns for for- migrants, who are also charac terized by eign-born Hispanics include positive health healthy selection at time of immigration. 9 selection at the time of immigration (that While longevity patterns for older His- is, healthier individuals are more likely to panics, particularly the foreign-born, are migrate) and favorable health-related be - clear ly favorable in comparison with havior, particularly low levels of cigarette whites, this is not the case in other health smoking.7 domains. For example, levels of physical Across national-origin groups, most disability among older Hispanics–both studies report higher life expectancy for U.S.- and foreign-born–are signi½cantly Hispanics who originate from Cuba and worse than those of whites at age sixty-½ve other countries in Central and South Amer- and above. Foreign-born Hispanic wom- ica, lower life expectancy among Puer to Ri - en have the highest level of physical disabil- cans, and life expectancy ½gures for Mex- ity when compared with U.S.-born whites, ican-origin Hispanics that are very similar blacks, and Hispanics and, thus, they are to those of all Hispanics. Lower life expec - characterized among these groups by the tancy (that is, higher mortality rates) for unique combination of the longest life ex - Puer to Rican–origin Hispanics have been pectancy coupled with the longest average attributed to their lower level of healthy period of time spent disabled.10 This is not immigrant selection (since Puerto Ricans to minimize the importance of the high lev- are U.S. citizens and easily migrate between els of physical disability also exhibited by Puerto Rico and the U.S. mainland) and U.S.-born Hispanic women and men, who their low socioeconomic status. Cubans, on share similar disability pro½les with black the other hand, are thought to be character - women and men. At older ages, both for- ized by healthier selection pro½les at time eign-born and U.S.-born Hispanic women of immigration and have long experienced also exhibit lower levels of physical func- a positive reception into the middle class of tioning (for example, gait speed and grip U.S. society.8 Low rates of smoking (and strength) than whites and similar levels as ligh ter smoking among those who do blacks.11

22 Dædalus, the Journal ofthe American Academy of Arts & Sciences Clearly, more research is needed to un- and respiratory disease mortality exhibit- Robert A. derstand why relatively long lives among ed by both foreign- and U.S.-born Hispan- Hummer & 15 Mark D. Hispanics are not coupled with low levels ics. Lower heart disease mortality among Hayward of disability and high levels of physical older Hispanics (particularly the foreign- functioning. One of our working hypoth - born) is also associated with their lower eses is that many Hispanics experience levels of cigarette smoking. pronounced “wear and tear” after years This brief overview clearly shows that, and years working in physically demand- at present, both Hispanic women and men ing occupations (such as in hospitality and have lower overall mortality rates and lon - food service, child care, domestic service, ger life expectancies than their more socio - construction, agriculture, and meat pro- economically advantaged white counter- cessing), placing them at disproportionate parts. The longer life-expectancy levels for risk of physical dif½culties and disability Hispanics are concentrated among the for - at older ages, even in the context of their eign-born and characterize most, but not relatively long lives. Long-term exposure all, of the national-origin Hispanic sub- to adverse socioeconomic conditions–for groups. Again, healthy immi grant selection example, higher rates of child poverty and to the United States and positive health lower levels of educational attainment behavior–particularly lower cigarette than whites–may also be partly responsi- consumption relative to whites–seem to ble for their disadvantaged patterns of dis - be the keys to this epidemiologic paradox ability and physical functioning. of longevity. For the Hispanic population, Hispanics also exhibit higher rates of this is all good news. Unfortunately, as we some, but not all, chronic morbidities in have shown, not all of the news is good. comparison with whites. Obesity, diabe- Perhaps most disturbing are the very high tes, and overall metabolic risk, for ex am - levels of disability and poor physical func - ple, are higher in most studies of middle- tioning among older Hispanics, which will aged and older Hispanics (both foreign- increas ingly challenge caregivers, families, and U.S.-born) compared to whites; some health agencies, and governments in a rap - (though not all) studies also document idly aging population. High levels of dia- somewhat higher levels of hypertension betes and metabolic risk among middle- among foreign- and U.S.-born Hispanics aged and older Hispanics, particularly the than among whites.12 U.S.-born Hispanics U.S.-born, also pose looming challenges. have also been shown to have moderately Our attention now turns to four of the key higher levels of inflammation risk com- health-related concerns facing the rapidly pared with whites.13 On the other hand, growing and aging Hispanic population. other morbidity rates are lower for Hispan- ics. Of greatest signi½cance, most studies The success or failure of U.S. immigra- ½nd a lower prevalence of cancer (with tion policy to effectively respond to the some site-speci½c exceptions) and lung dis - roughly 11 million undocumented residents ease among both foreign-born and U.S.- (approximately three-fourths of whom born Hispanics compared with whites– are Hispanic) will have long-term conse- patterns consistent with Hispanics’ histor - quences for health and longevity patterns ically much lower levels of cigarette con - among older Hispanics.16 In an important sump tion.14 Lower cancer- and lung dis- essay in the Summer 2013 issue of this jour- ease–related morbidity among Hispanics nal, sociologist Douglas Massey argued relative to whites aligns with the signi½ - that rapid growth in the undocumented cantly lower all-site cancer, lung cancer, population between 1965 and 2008 was an

144 (2) Spring 2015 23 Hispanic unintended consequence of U.S. immi- and social services that characterize many Older Adult gration and border control policies in con- undocumented immigrants will undoubt- Health & Longevity junction with the simultaneous increase edly have negative health consequences for in the in the economic and social integration of this segment of the U.S. population in the United States Mexico and the United States. He further decades to come. argued that a path to citizenship for un- Second, policies focused on undocu- documented residents of the United States mented immigrants are very likely to have is a necessary and critical step toward the important spillover effects on children of future social and economic well-being of immigrants as well as on the Hispanic com - Hispanics in U.S. society.17 munity as a whole. An estimated 73 percent Massey’s arguments are directly rele- of children of undocumented immigrants vant to the future health and longevity of are U.S.-born citizens; thus, the future the aging Hispanic population in at least health and well-being of these children will two important ways. First, the growth in also be in part dependent upon the resolu - the undocumented immigrant population of their parents’ legal status.19 While tion means that a signi½cant portion of data are not available to assess the relation- the U.S. population is largely invisible in ship between parents’ legal status and chil- the country’s data monitoring systems. dren’s well-being, there is a substantial Thus, there are no credible estimates of body of evidence pointing to the perni- mortality rates, disability levels, or mor- cious effects of poverty and fam ily stress bidity patterns at the national level for this for children’s long-term health. Thus, the sizable subgroup (roughly 3.5 percent) of health of children of undocumented im - the U.S. population. An unknown number migrants is in a very real sense the embod- of undocumented residents may be in - iment of parents who live with substantial cluded in health surveys, census records, uncertainty and stress and who lack access and vital statistics, but it is dif½cult or im- to basic social services, health care, and le - possible to know with any certainty how gal rights. More generally, the issue of im - their health patterns compare to other sub- migrant legal status has been a critical one groups or to the population as a whole. in the Hispanic community for decades, Despite their statistical invisibility, there and the intensity of the debate has only in - are reasons to suspect that the health and creased in the fourteen years following the longevity patterns for undocumented 9/11 terrorist attacks, given stricter U.S. immigrants are not very good. For exam- bor der controls. A positive solu tion to the ple, estimates from the Pew Hispanic Cen- legal status of the United States’ 11 million ter show (not surprisingly) that undocu- undocumented immigrants, 8 million of mented immigrants have substantially whom are Hispanic, would be critical to all lower household incomes and lower levels Hispanics achieving the full integration of insurance coverage than the U.S.-born that is funda mental to their long-term resident population.18 Moreover, while the health and well-being. A federal policy that vast majority of undocumented Hispanic encour ages legalization of undoc umented residents of the United States are currently immigrants in a humane and healthy way between the ages of twenty and ½fty, the will substantially strengthen the prospects clock is ticking and these undocumented of favorable future health and longevity adult immigrants will move into old age patterns among older Hispanics. near mid-century. The poor wages, harsh working conditions, high levels of stress Asecond major health concern for His- and fear, and lack of access to health care panics stems from the exclusion of undoc-

24 Dædalus, the Journal ofthe American Academy of Arts & Sciences ument ed immigrants from purchasing sage of the aca. A key structural barrier Robert A. health in surance under the provisions of to health insurance coverage among both Hummer & Mark D. the 2010 Patient Protection and Affordable the legal immigrant and U.S.-born seg- Hayward Care Act (aca).20 This results in signi½ - ments of the Hispanic population is the cant barriers for undocumented immi- limited eligibility for low-income adults grants seeking health services–including of all racial/ethnic groups to qualify for emergency care–compared to both legal Medicaid coverage in states that thus far immigrants and those born in the United have decided not to expand Medicaid cov - States. Undocumented immigrants are less erage under the aca (these include Texas, likely to seek care, and when they do, they Florida, Georgia, and twenty-one oth- are at greater risk of presenting more ad - ers).24 As with immigration policy toward vanced and complicated health problems undocumented residents, this policy issue and developing functional limitations and may not necessarily be harming the over- disabilities downstream.21 Undocument- all health and longevity patterns of cur- ed status and lack of access to health in - rent older-aged Hispanics, the vast majori- surance coverage are thus inextricably ty of whom are legal residents of the United linked with the aca for 8 million or so States and have health insurance coverage Hispanics, and this link has direct rami½ - through Medicare. However, it is critical cations for Hispanic health far into the to keep in mind that the older-aged Hispan - future. ic population of tomorrow is the working- Compounding this situation is the al - aged Hispanic population of to day; this ready low level of health insurance cover- group’s current health insurance coverage age for the Hispanic population as a whole: may help determine whether they are the U.S. Census Bureau estimated that 29.1 healthy and disease-free in older adult hood percent of the Hispanic population (15.5 or whether, on the contrary, they develop million) was uninsured in 2012 (down from conditions that could have been treated over 30 percent in 2011), the highest pro- much earlier in life or avoided altogether. portion by far of any ethnic group in the country.22 Thus, even if every single un - The third major health concern for His- documented Hispanic resident of the panics is rooted in two key health-related United States suddenly purchased health behaviors: smoking and the combination insurance coverage, there would still be of poor nutrition and low physical activity at least 7.5 million uninsured Hispanic res- that results in obesity. Smoking and obesi- idents in the United States–around 15 per- ty have long been identi½ed as the two most cent of the Hispanic population. This 7.5- important behavior-related causes of million ½gure includes both legal Hispanic poor health and premature death in the immigrants as well as U.S.-born Hispanics. United States. National trends in obesity A 2014 Gallup Poll provides estimates and smoking point to the potential ero- of racial/ethnic differences in aca uptake sion of Hispanics’ current advantage in life among the previously uninsured popula- expectancy relative to whites and a wid - tion.23 Less than 11 percent of uninsured ening of Hispanics’ existing disadvantages Hispanics obtained coverage during the in old-age disability and physical function- open-enrollment period, compared with ing. With respect to obesity, recent stud- 16 percent increases for blacks and 14 per- ies have projected that a long-term increase cent for whites. This is a worrisome pattern in the prevalence of obesity will reduce given the overall lower level of insurance U.S. life expectancy in the future.25 Al- coverage among Hispanics prior to the pas- though none of this work has directly

144 (2) Spring 2015 25 Hispanic examined Hispanics, it is very likely that white and black women’s smoking levels Older Adult Hispanics will disproportionately bear the have exhibited rapid declines in re cent de - Health & Longevity brunt of this effect. Obesity prevalence cades while Hispanic women’s levels have in the has increased precipitously in countries remained about the same, suggesting that United States from which many Hispanic immigrants white and black women’s lung cancer rates originate, particularly Mexico,26 and it has will also soon begin to fall while Hispanic increased disproportionately among His- women’s rates will not.31 Thus, on the panics in the United States.27 Given the whole, the advantage relative to other high-calorie, high-fat diets that have be - groups that U.S. Hispanics have from come characteristic of the United States smoking less may be disappearing. With and Mexico in the last several decades, di - the disproportionate and precipitous rise etary behavior will need to change quickly in obesity among Hispanics compared to and in dramatic ways for adverse health and whites, then, Hispanics’ current life-ex - longevity consequences among Hispanics pectancy advantage may well become a dis- to be minimized. As emphasized above, it advantage in the not-too-distant future. is not necessarily current older-aged His- panics who are at particular risk; rather, The fourth concerning factor for future the much larger group of working-age old-age Hispanic health and longevity pat- Hispanics (who will become tomorrow’s terns is their continued disadvantaged so- elderly Hispanics) are facing this growing cioeconomic pro½le. One might ask why threat. this is concerning, given that current His- On the upside, recent work also shows panic old-age mortality rates and life ex - that future declines in overall U.S. life ex - pectancy levels are favorable compared pec tancy due to increasing obesity may to whites (as reviewed above), despite the be counterbalanced by life expectancy in- overall low socioeconomic status of the creases resulting from decreases in smok- Hispanic population. After all, does not the ing.28 That said, the relative balance of combination of relatively high life expec - obesity and smoking effects could differ tancy despite socioeconomic disadvantage substantially for Hispanics in comparison de½ne the epidemiologic paradox? Two with the U.S. population as a whole. His- points are worth noting. First, our recent panics have smoked less in the past and work demonstrates that current older-aged continue to smoke less than either whites Hispanics would have even lower mortal- or blacks. This lower level of smoking is ity rates and higher life expectancies than responsible for up to one-half of Hispan- they already have if their levels of educa- ics’ current life expectancy advantage over tion and income were similar to those of whites.29 However, this smoking-related whites.32 In other words, Hispanic longev- advantage for Hispanics may be diminish- ity patterns are not immune to low socio - ing as blacks and whites catch up. Lung economic status. Second, socioeconomic cancer incidence and mortality rates for re sources are perhaps more fundamen- white and black men have declined pre- tally important for individual-level health cipitously since 1999, while the same rates in the United States than ever before in for Hispanic men have declined only mod - our nation’s history.33 Characteristics such estly.30 Women’s lung cancer death rates as educational attainment, stable employ - for all racial/ethnic groups have remained ment, decent income, and wealth holdings stable since the late 1990s, with the rate provide individuals with access to flexi- for Hispanics considerably lower than ble resources–goods, information, tech- both whites and blacks; however, both no logies, social ties, and psycho-social re -

26 Dædalus, the Journal ofthe American Academy of Arts & Sciences sources–that are particularly important groups erode when comparing the immi- Robert A. for negotiating health-related behavior and grant generation to subsequent genera- Hummer & 39 Mark D. care in the hyper-competitive and in for - tions. Socioeconomic disadvantages Hayward mation-based twenty-½rst century. among U.S.-born Hispanics may play an Looking ahead, there is no logical reason important role in these generational why Hispanic health outcomes will not be changes in mortality patterns. Given the influenced by socioeconomic status, as well-established and strengthening rela- they are in other racial/ethnic groups. For tionship between socioeconomic status the country as a whole, life expectancy has and health in the United States, Hispanic risen disproportionately among persons health and longevity patterns may soon with a college-level education or higher, lag behind those of whites if aggressive while less-educated adults have actually ex- efforts are not undertaken to enhance so - perienced a decrease in life ex pectancy in cioeconomic achievement for all disadvan- recent decades.34 Yet unfortunately, His- taged groups. Policy efforts must focus panic adults currently have the lowest rate on making advanced education (and of college (and high school) graduation out higher-quality education at all levels) of all other U.S. racial/ethnic groups.35 It much more attainable for socially disad- is not surprising, then, that Hispanics are vantaged groups; early-educational inter- concentrated in low-wage service-sector ventions may also be particularly impor- jobs. In addition, Hispanic households tant and fruitful for children who have hold, on average, less than 6 percent of the Span ish-speaking immigrant parents. In - wealth that white households do; this gap come policies–minimum wage, child care, has widened consider ably in recent years.36 paid family leave, housing subsidies, and Hispanic socioeconom ic disadvantages rel- more–can also work to improve the so cio - ative to whites per sist even to the second- economic status of families who are de- and third- genera tion U.S.-born Hispan- pendent upon low-wage service work. ics.37 Such pro nounced socioeconomic There is little doubt that future Hispanic dis advantages could have damaging effects health and longevity patterns will be in - on Hispanics’ health and longevity pat- fluenced by how policy-makers help shape terns in the coming decades. the social and economic future of Amer- Aggressive policies are needed to deal ica’s socioeconomically disadvantaged with the substantial disadvantages in ed - populations. ucational attainment, income, and wealth experienced by these sizable segments of While current longevity patterns for the U.S. population. Such socioeconomic Hispanics relative to whites are favorable, policies will also act as health policies: old-age functioning and disability patterns more socioeconomic empowerment will for Hispanics compare unfavorably. Pat- give individuals more access to care and terns for some morbidity conditions– better tools to make health-related deci- most notably diabetes and metabolic issues sions.38 Although we have focused on His- –place Hispanics at higher risk than other panics, this issue is germane to all socio - U.S. racial/ethnic groups. Future His pan ic economically disadvantaged groups. But health and longevity patterns are troubling since Hispanics will constitute such a large for a number of reasons: the unresolved percentage of our aging population, their legal status of over 8 million unauthorized status will disproportionately influence the immigrants, continued low levels of His- future. Currently, it is clear that Hispanic panic insurance coverage even after health mortality advantages relative to other care reform, some unfavorable trends in

144 (2) Spring 2015 27 Hispanic Hispanic health behavior, and continued, of policies to address these issues will be Older Adult substantial socioeconomic disadvantages of particular value to the large and grow- Health & Longevity for Hispanics relative to whites. While ing elderly Hispanic population, and by in the these risks and vulnerabilities are shared extension, the U.S. elderly population in United States with other groups as well, development general.

endnotes * Contributor Biographies: ROBERT A. HUMMER is Centennial Commission Professor in the Lib- eral Arts in the Department of Sociology and Faculty Research Associate in the Population Re - search Center at the University of Texas at Austin. His research focuses on the accurate descrip - tion and more complete understanding of health and mortality disparities in the U.S. population. He is the coauthor of Living and Dying in the U.S.A.: Health, Behavioral, and Social Differentials of Adult Mortality (with Richard Rogers and Charles Nam, 2000) and the coeditor of the volume Religion, Families, and Health: Population-Based Research in the United States (with C. E. Ellison, 2010). He has also published extensively in demography, sociology, and public health journals.

MARK D. HAYWARD is Professor of Sociology, Centennial Commis sion Professor in the Lib- eral Arts, and Faculty Research Associate in the Population Re search Center at the Univer- sity of Texas at Austin. His primary research focus is how life-course exposures and events influence the morbidity and mortality experiences of the older population. He has published extensively on population health topics in such jour nals as Demography, Journal of Health and Social Behavior, and Social Science & Medicine.

Authors’ Note: Funding was provided by the MacArthur Foundation Research Network on an Aging Society (John W. Rowe, Chair), and by infrastructural research support (5 R24 HD042849) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development to the Population Research Center at the University of Texas at Austin.

1 Karen R. Humes, Nicholas A. Jones, and Roberto R. Ramirez, Overview of Race and Hispanic Origin: 2010 (Washington, D.C.: U.S. Census Bureau, 2011). 2 U.S. Census Bureau, “Table 6. Percent of the Population by Race and Hispanic Origin for the United States: 2015 to 2060,” 2012 National Population Projections (Washington, D.C.: U.S. Census Bureau, Population Division, 2012), https://www.census.gov/population/projections/ data/national/2012.html. 3 Jennifer M. Ortman, Victoria A. Velkoff, and Howard Hogan, An Aging Nation: The Older Popula- tion in the United States (Washington, D.C.: U.S. Census Bureau, 2014). 4 Kyriakos S. Markides and Jeanne Coreil, “The Health of Southwestern Hispanics: An Epi- demiologic Paradox,” Public Health Reports 101 (1986): 253–265; and Kyriakos S. Markides and Karl Eschbach, “Hispanic Paradox in Adult Mortality in the United States,” in Interna- tional Handbook of Adult Mortality, ed. Richard G. Rogers and Eileen M. Crimmins (New York: Springer, 2011), 227–240. 5 Elizabeth Arias, “United States Life Tables by Hispanic Origin,” Vital and Health Statistics Series 2, No. 152 (Hyattsville, Md.: National Center for Health Statistics, 2010), 1–35; Irma T. Elo, Cassio M. Turra, Bert Kestenbaum, and B. Renee Ferguson, “Mortality Among Elderly Hispanics in the United States: Past Evidence and New Results,” Demography 41 (2004): 109 –128; Joseph T. Lariscy, Robert A. Hummer, and Mark D. Hayward, “Hispanic Older Adult Mortality in the United States: New Estimates and An Assessment of Factors Shaping the Hispanic Paradox,” Demography 52 (1) (2015). 6 Lariscy, Hummer, and Hayward, “Hispanic Older Adult Mortality in the United States.”

28 Dædalus, the Journal ofthe American Academy of Arts & Sciences 7 Andrew Fenelon, “Revisiting the Hispanic Mortality Advantage in the United States: The Role Robert A. of Smoking,” Social Science and Medicine 82 (2013): 1–9; and Gopal K. Singh and Robert A. Hiatt, Hummer & “Trends and Disparities in Socioeconomic and Behavioural Characteristics, Life Expec - Mark D. tancy, and Cause-Speci½c Mortality of Native-Born and Foreign-Born Populations in the Hayward United States, 1979–2003,” International Journal of Epidemiology 35 (2006): 903–919. 8 Robert A. Hummer, Richard G. Rogers, Sarit H. Amir, Douglas Forbes, and W. Parker Frisbie, “Adult Mortality Differentials among Hispanic Subgroups and Non-Hispanic Whites,” Social Science Quarterly 81 (2000): 459–476. 9 Fenelon, “Revisiting the Hispanic Mortality Advantage in the United States”; and Markides and Eschbach, “Hispanic Paradox in Adult Mortality in the United States.” 10 Karl S. Eschbach, Soham Al-Snih, Kyriakos S. Markides, and James S. Goodwin, “Disability and Active Life Expectancy of Older U.S.- and Foreign-Born Mexican Americans,” in The Health of Aging Hispanics, ed. Jacqueline L. Angel and Keith E. Whit½eld (New York: Springer, 2007): 40–49; and Mark D. Hayward, Robert A. Hummer, Chi-Tsun Chiu, César González-González, and Rebeca Wong, “Does the Hispanic Paradox in U.S. Adult Mortality Extend to Disability?” Population Research and Policy Review 33 (2014): 81–96. 11 Steven A. Haas, Patrick M. Krueger, and Leah Rohlfsen, “Race/Ethnic and Nativity Disparities in Later Life Physical Performance: The Role of Health and Socioeconomic Status Over the Life Course,” The Journals of Gerontology: Psychological Sciences & Social Sciences 67 (2012): 238–248. 12 Eileen M. Crimmins, Jung Ki Kim, Dawn E. Alley, Arun Karlamangia, and Teresa Seeman, d “Hispanic Paradox in Biological Risk Pro½les,” American Journal of Public Health 97 (2007): d 1305–1310; and Gopal K. Singh and Mohammad Siahpush, “Ethnic-Immigrant Differentials in Health Behaviors, Morbidity, and Cause-Speci½c Mortality in the United States: An Analysis of Two National Data Bases,” Human Biology 74 (2002): 83–109. For statistics on hypertension, see Crimmins et al., “Hispanic Paradox in Biological Risk Pro½les”; and Zhenmei Zhang, Mark D. Hayward, and Chuntian Lu, “Is There a Hispanic Epidemiologic Paradox in Later Life? t A Closer Look at Chronic Morbidity,” Research on Aging 34 (2012): 548–571. 13 Crimmins et al., “Hispanic Paradox in Biological Risk Pro½les.” c 14 Zhang et al., “Is There a Hispanic Epidemiologic Paradox in Later Life?” 15 Lariscy, Hummer, and Hayward, “Hispanic Older Adult Mortality in the United States.” 16 Jeffrey S. Passel and D’Vera Cohn, A Portrait of Unauthorized Immigrants in the United States (Washington, D.C.: Pew Hispanic Center, 2009). 17 Douglas S. Massey, “America’s Immigration Policy Fiasco: Learning from Past Mistakes,” Dædalus 142 (3) (Summer 2013): 5–15. 18 Passell and Cohn, “A Portrait of Unauthorized Immigrants to the United States.” 19 Massey, “America’s Immigration Policy Fiasco.” 20Steven P. Wallace, Jacqueline Torres, Tabahir Sadegh-Nobari, Nadereh Pourat, and E. Richard Brown, “Undocumented Immigrants and Health Care Reform,” Final Report to The Common - wealth Fund, ucla Center for Health Policy Research, 2012. 21 Ibid.; and Marc L. Berk, Claudia L. Schur, Leo R. Chavez, and Martin Frankel, “Health Care Use Among Undocumented Latino Immigrants,” Health Affairs 19 (2000): 44–57. 22 y Carmen DeNavas-Walt, Bernadette D. Proctor, and Jessica C. Smith, “Income, Poverty, and Health Insurance Coverage in the United States: 2012,” Current Population Reports P60-245 t (Washington, D.C.: U.S. Census Bureau, 2013), 1–77. 23 Jenna Levy, “U.S. Uninsured Rate Drops to 13.4%,” Gallup, 2014, http://www.gallup.com/poll/ 168821/uninsured-rate-drops.aspx (accessed September 4, 2014).

144 (2) Spring 2015 29 Hispanic 24 Henry J. Kaiser Family Foundation, “Health Coverage for the Hispanic Population Today and Older Adult Under the Affordable Care Act,” Kaiser Commission of Medicaid and the Uninsured: Key Facts Health & (Washington, D.C.: Kaiser Family Foundation, 2013): 1–6, http://www.kff.org/kcmu. Longevity in the 25 Samuel H. Preston, Andrew Stokes, Neil K. Mehta, and Bochen Cao, “Projecting the Effect United of Changes in Smoking and Obesity on Future Life Expectancy in the United States,” Demog- States raphy 51 (2014): 27–49. 26 Juan A. Rivera, Simon Barquera, Teresa González-Cossío, Gustavo Olaiz, and Jaime Sepulveda, “Nutrition Transition in Mexico and in Other Latin American Countries,” Nutrition Reviews 62 (2004): S149–S157. 27 Katherine M. Flegal, Margaret D. Carroll, Brian K. Kit, and Cynthia L. Ogden, “Prevalence of Obesity and Trends in the Distribution of Body Mass Index among U.S. Adults, 1999–2010,” JAMA: The Journal of the American Medical Association 307 (2012): 491–497. 28 Preston, Stokes, Mehta, and Cao, “Projecting the Effect of Changes in Smoking and Obesity on Future Life Expectancy in the United States.” 29 Andrew Fenelon, “Revisiting the Hispanic Mortality Advantage in the United States: The Role of Smoking,” Social Science and Medicine 82 (2013): 1–9; and Lariscy, Hummer, and Hayward, “Hispanic Older Adult Mortality in the United States.” 30 Centers for Disease Control and Prevention, “Lung Cancer Rates by Race and Ethnicity,” http://www.cdc.gov/cancer/lung/statistics/race.htm. 31 American Lung Association Research and Program Service (Epidemiology and Statistics Unit), “Trends in Tobacco Use,” 2011, http://www.lung.org/½nding-cures/our-research/trend -reports/Tobacco-Trend-Report.pdf 32 Lariscy, Hummer, and Hayward, “Hispanic Older Adult Mortality in the United States.” 33 Bruce G. Link and Jo Phelan, “Social Conditions as Fundamental Causes of Disease,” Journal of Health and Social Behavior 36 (extra issue) (1995): 80–94; and Jo C. Phelan, Bruce G. Link, and Parisa Tehranifar, “Social Conditions as Fundamental Causes of Health Inequalities: The- ory, Evidence, and Policy Implications,” Journal of Health and Social Behavior 51 (extra issue) (2010): S28–S40. 34 S. Jay Olshansky, Toni Antonucci, Lisa Berkman, Robert H. Binstock, Axel Boersch-Supan, John T. Cacioppo, Bruce A. Carnes, Laura L. Carstensen, Linda P. Fried, Dana P. Goldman, James Jackson, Martin Kohli, John Rother, Yuhui Zheng, and John Rowe, “Differences in Life Expectancy Due to Race and Educational Differences Are Widening, and Many May Not Catch Up,” Health Affairs 31 (8) (2012): 1803–1813. 35 Marta Tienda and Faith Mitchell, eds., Multiple Origins, Uncertain Destinies: Hispanics and the American Future (Washington, D.C.: The National Academies Press, 2006). 36 Rakesh Kochhar, Richard Fry, and Paul Taylor, Wealth Gaps Rise to Record Highs Between Whites, Blacks, and Hispanics (Washington, D.C.: Pew Research Center, 2011), http://www.pewsocial trends.org/½les/2011/07/sdt-Wealth-Report_7-26-11_final.pdf. 37 Tienda and Mitchell, Multiple Origins, Uncertain Destinies. 38 Robert F. Schoeni, James S. House, George A. Kaplan, and Harold Pollack, eds., Making Ameri- cans Healthier: Social and Economic Policy as Health Policy (New York: Russell Sage Foundation, 2008). 39 Lariscy, Hummer, and Hayward, “Hispanic Older Adult Mortality in the United States”; and Fernando Riosmena, Bethany G. Everett, Richard G. Rogers, and Jeff A. Dennis, “Negative Acculturation and Nothing More? Cumulative Disadvantage and Mortality during the Immi- grant Adaptation Process among Latinos in the U.S.,” International Migration Review (forthcoming 2015).

30 Dædalus, the Journal ofthe American Academy of Arts & Sciences d The Future of Intergenerational Relations t in Aging Societies

Frank F. Furstenberg, Caroline Sten Hartnett, Martin Kohli & Julie M. Zissimopoulos f y Abstract: As the pressure mounts to reduce the public costs of supporting rapidly aging societies, responsibility for supporting elderly people will increasingly fall on their family members. This essay explores the family’s capacity to respond to these growing challenges. In particular, we examine how family change and growing inequality pose special problems in developed nations, especially the United States. This essay mentions a series of studies supported by the MacArthur Foundation Research Network on an Aging Society that aim to examine the future of intergenerational exchange. We focus particularly on adults who have dependent and young-adult children and who must also care for elderly parents, a fraction of the population that will grow substantially in the coming twenty-½ve years. d

f FRANK F FURSTENBERG Family systems are among the most enduring and . , a Fellow universal institutions that human beings have de - of the American Academy since 1994, is the Zellerbach Family Pro- vised; in one form or another, they exist in every so - fessor of Sociology Emeritus at the ciety as a social arrangement for regulating repro- University of Pennsylvania. duction, supporting the young, and caring for the elderly. The survival of family systems stems from CAROLINE STEN HARTNETT is an Assistant Professor of Sociology at their remarkable ability to transform their struc- t the University of South Carolina. ture and practices in response to new demographic, MARTIN KOHLI social, and economic conditions. That the family sys- is Distinguished tem has persevered despite the vast global changes e Bremen Professor at the Bremen In - ternational Graduate School of So - of the last century attests to this resiliency and adapt- cial Sciences and Professor of So- ability. But there are a number of indications that the l ciology, Emeritus, at the European family’s capacity to perform its traditional functions University Institute, Florence. is becoming strained as the pace of social and eco- JULIE M. ZISSIMOPOULOS is an nomic change quickens around the globe. Assistant Professor at the Sol Price Family formation and childbearing, robust a half- School of Public Policy and the As - century ago, have slowed to below population- sociate Director of the Schaeffer replacement levels in most developed nations. Mar- d Center at the University of South- riage has become a less central and stable institution, ern California. making child care more precarious. The breakdown (*See endnotes for complete contributor of the gender-based division of labor has challenged biographies.) women and men to develop more complex routines

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00328 31 The Future for managing work and family roles. Great- only increase the burden on working-age of Intergen- er parental investment and support is re - adults already challenged by the temporal erational Relations quired for childrearing today: lengthier and ½nancial demands of supporting the in Aging and increasingly expensive post-second- elderly while simultaneously having to in - Societies ary education and depressed job prospects vest more in caring for and educating the have prolonged young adults’ dependency young. In this essay, we explore whether on their parents. On the other hand, a rapid families will be up to this task. increase in longevity has enlarged the We begin this inquiry by briefly sum- ranks of the elderly, contributing–along marizing some salient ½ndings on prac- with the smaller number of the young– tices of intergenerational support and ex- to an increase in their share of the total change in economically developed nations. population. This adds to the elderly’s ca - We will then discuss changing conditions pacity to assist younger generations, that are likely to complicate (and possibly though at the same time it increases the alter or erode) these existing familial re- future burdens on younger generations to lationships and patterns of intergenera- support and care for elderly family mem- tional resource allocation. We also exam- bers when they become frail and incapac- ine coping mechanisms families develop itated. to accommodate intergenerational needs. The decline of fertility and growth of In the second part of the paper, we will de- longevity have produced a rapidly aging scribe the work of the MacArthur Founda - society in many parts of the world. This tion Research Network on an Aging Soci- pattern of population aging will be accen - ety, which has devised a series of projects tuated as the large cohort of individuals to explore these social issues linked with born after World War II–the baby boom - aging societies, and whose work may con - ers–reaches seniority. The demographic tribute signi½cantly to the knowledge base pro½le for each wealthy nation differs that informs U.S. policy on aging. We con- depending on the timing of its baby boom clude by considering some possible dilem - and the magnitude of its baby bust (a mas that policy-makers may face in their consequence of rapidly decreasing fertility efforts to support families. in the latter third of the last century). Some are aging sooner or more rapidly than oth - A vast amount of research has been done ers; these nations have provided some in- over the past several decades on how gen- sights about the effects of population aging. erations within family systems support (For more on demographic shifts in the and assist each other over time. Patterns United States, see S. Jay Olshansky’s essay of intergenerational exchange operate in this volume.) quite differently in traditional societies, Public systems such as Social Security developing countries, and advanced econ - and Medicare were devised a century or omies. Traditional societies have often de - more ago partly to relieve some of the eco - vised means of protecting the elderly nomic and social pressures on the family generation (G1) by maintaining intergen- –midlife parents in particular–to assume erational households or mandating that complete responsibility for the care of both the middle generation (G2) provide ½nan - the young and the elderly. Due to popula- cial support to elderly parents, though old - tion aging, these public support systems er parents (G1) may also continue to assist will face considerable pressure in many their adult children by caring for their off - so cieties, including in the United States. spring (G3) or working. Two distinctive Any reductions in public expenditure will features of traditional systems are that the

32 Dædalus, the Journal ofthe American Academy of Arts & Sciences oldest generation 1) does not survive in parents when they become frail or inca- Frank F. great numbers and 2) produces a relative ly pacitated, but are less often called upon to Furstenberg, Caroline Sten high number of children who can provide provide economic assistance to elderly par- Hartnett, Martin support as they come of age. Moreover, low ents (many of whom enjoy the economic Kohli & Julie M. geographical mobility tends to keep fam - security provided by their social security Zissimopoulos ilies in the same locale, ensuring that par- bene½ts, pensions, and savings). Those in ents will have surviving children around need of personal care often ½rst turn to to provide support when they are no longer spouses or partners, but children may also able to take care of themselves.1 step in, either as caregivers or as organizers This general pattern of an upward flow and monitors of third-party or institution- of resources from the middle to the oldest al care. generation typically persists in developing Admittedly, this is an oversimpli½ed pic- countries, though due to the expansion of ture of intergenerational exchange in labor markets, second-generation mem- wealthy countries; there is much variation bers often move away from their birth com - both within and across individual nations. munities to seek higher incomes. In inter- First, the magnitude of public support national migration, this may lead to con- varies greatly among welfare systems. Sec- siderable remittances from G2s to G1s. On ond, countries also differ in terms of the the other hand, children may sometimes forms of their support systems (family- have to be left behind in the care of their based versus individualized).4 Third, the grandparents, especially when mothers flow of resources works differently across move to wealthier countries (often to different economic strata. Among the poor become paid care workers for children or and near-poor, children may be called up- the elderly). The introduction of old-age on to help out both economically and non - pensions–frequently the ½rst step of pub - monetarily throughout their lives and can - lic social assistance–greatly improves the not readily afford to purchase assistance ½nancial position of the elderly and may for older relatives, as might be common lead to a reversal of the ½nancial flow by among the more privileged. Fourth, there enabling them to aid the families of their is a great deal of evidence showing that ra- adult children and help cover the school- cial and ethnic differences correlate with ing expenses of their grandchildren.2 dif ferent patterns of support and ex - In nations with advanced economies, the change.5 Some ethnic groups are much flow of economic assistance more gener- more likely to maintain intergenerational ally moves downward. Children and grand - households in which the volume of support children require more and longer support and assistance is more intense and recip- due to extended and more costly educa- rocal over time.6 (Interestingly, however, tion and dif½cult transitions into work; there is a paucity of evidence on the work- conversely, the elderly are often econom- ings of intergenerational households in ically better-off thanks to public support, con temporary societies.) Finally, divorce savings, and improvements in health care. and remarriage affect the amount of sup- In all European nations, parents continue port provided both downward and upward to provide more economic assistance to throughout the life course. Divorced fa- their children and grandchildren than they thers not infrequently lose contact with receive (at least through their seventies). some or all of their children, who in turn This is true even if we include nonmone- are unwilling or unavailable to as sist their tary help.3 Children (and possibly grand- fathers in later life. Stepfathers may ½ll in children) often reciprocate by aiding their to some extent, but there is strong evidence

144 (2) Spring 2015 33 The Future that children feel less ob ligated to care for similar to that faced by young parents. At of Intergen- elderly stepparents than parents.7 the same time, unless current patterns of erational Relations As we contemplate a potential retreat of family formation rapidly reverse–and in Aging public systems of assistance for the elderly there is no reason to expect that they will– Societies in the next several decades, there are sound a rising number of young adults will not reasons to be concerned about families’ begin having children until their thirties ability to pick up the slack. This is espe- or even later. More than one-third of births cially true in the United States, where the are now to women over thirty, and the population is becoming more economi- average age at ½rst birth has risen steadily, cally strati½ed, more ethnically diverse, and from twenty-one in 1970 to nearly twenty- more varied in its family forms. six in 2012.11 Not all of these young adults remain dependents while forming partner- Impending demographic change in the ships and having children, but many do size of elderly and middle-aged cohorts receive partial or complete support from drives much of the concern about the pres- their parents in their early adult years. The sures likely to mount on parents in midlife proportion of young adults being assisted and late midlife (roughly ages forty-½ve by their parents has been steadily rising to sixty-four).8 These middle-aged parents over the past several decades.12 Were this (mothers especially) are often called the pattern of the prolongation of early adult - “sandwiched generation,” squeezed by the hood to continue, it would result in an even simultaneous demands of supporting their greater level of “sandwiching,” mandating children, assisting a frail elderly parent, and that a growing proportion of midlife adults being in the labor force. Some ar gue that divide time and resources between their this term exaggerates the pressure faced by parents, their children, and possibly even this generation, because many parents take their grandchildren. active steps to avoid being in this position These demographic trends of greater or do not have to ½ll all three roles at once. longevity and later family-formation are Nonetheless, it is highly likely that a sub- compounded by rapid changes in marriage stantial percentage of families will face practices that have tended to make family these circumstances in the next two or obligations more uncertain. Compared to three decades. Though there is little re- four or ½ve decades ago, young adults mar- search describing the incidence and dura- ry later and divorce more often, making co- tion of episodes of being “sandwiched” for habitation and remarriage more common middle-aged adults, we know that about and socially acceptable. Marriage has be- one-third of women in late middle age (become a less central institution for organiz- tween ½fty-½ve and sixty-nine) report giving childbearing than it was a half-cen tury ing support to both children and parents at ago; a growing proportion of chil d ren are some point.9 This is likely to increase, since born out of wedlock, often into cohabiting the proportion of adults who have both unions (which are less stable family forms, children and living parents is rising.10 particularly in the United States).13 The number of elderly (including par- We know all too little about felt obliga- ents, parents-in-law, and stepparents) who tions to extended kin in varied family cir- survive to become incapacitated or frail cumstances. Existing evidence suggests will increase as the baby-boom cohort that marriage–especially a ½rst marriage– advances into their later years in the next confers a greater sense of obligation to in- several decades. This poses a new challenge laws than do consensual unions.14 Cohab - for the reconciliation of family and work, itation may offer a more discretionary and

34 Dædalus, the Journal ofthe American Academy of Arts & Sciences conditional commitment to both the part- con tact between fathers and children.19 Frank F. ners and their extended kin. The growing Fam ily units are more complex, often con - Furstenberg, Caroline Sten number of remarriages and cohabitations tain ing step- or surrogate parents and half- Hartnett, Martin after a ½rst marriage may result in weak er siblings.20 Coresidence among three gen- Kohli & Julie M. obligations to in-laws and quasi–in-laws, erations is a common adaptive strategy Zissimopoulos reducing the level of support to in-laws for managing economic pressures and the that accompanied marriages in the past. greater need for assistance by kin.21 At As mentioned earlier, an ever-greater the same time, the elderly have signi½cant- fraction of children are growing up with ly shorter life spans, suggesting that pres- sporadic or limited contact with their bio - sures on midlife parents (often matri archs logical fathers (and a smaller but rising charged with keeping the family together) number instead have limited contact with are considerable. Cutbacks in public sup- their biological mothers, or have little con - port most acutely affect these overbur- tact with either parent). As stated previ- dened families, and it is unclear how par- ously, the evidence strongly suggests that ents will be able to invest adequately in children with limited contact with their their children while supporting frail par- fathers are–not unsurprisingly–far less ents or grandparents (shorter generations likely to provide assistance to them when result in more four-generation families in they become elderly.15 A third trend asso- this segment of the population, owing to ciated with family change may also affect earlier childbearing ages). the availability of support for older people Issues facing the privileged are likely to in the future: childlessness has been rising be less acute, if only because of their greater rapidly in many nations with advanced resources for purchasing assistance when economies, including the United States, needed. Nonetheless, many will face an where close to one-sixth of all women in overlap between demands from children their forties are currently childless.16 For- and from elder parents and in-laws. Af - tunately, much of the research on this topic fluent parents tend to provide high levels of suggests that the childless are not neces- support to young adult children and at sarily stranded in old age because many the same time have parents with relatively acquire surrogate caregivers who may or high life expectancies.22 As greater num - may not be kin.17 Nonetheless, the relative bers of baby-boomers reach old age, afflu- shortage of children among people enter- ent parents will have to contend with the ing midlife may eventually reduce the declining health of their parents and in- amount of assistance and care available to laws while they are still launching their certain subsets of the elderly. young-adult children. The constellation of family changes has Families of modest and moderate means not been uniformly distributed; markedly –the middle third of the economic distri- different family forms exist across the so - bution–face a different situation, at least cial strata. As inequality has widened in the in countries with limited social safety United States and elsewhere, family prac- nets such as the United States. They are tices among the privileged, the disadvan- squeezed economically because, though taged, and those in the middle have become they are ineligible for public assistance, more pronounced and distinct from one they may have obligations to ½nancially another.18 as sist their older parents and adult chil - Disadvantaged families are generally dren (who may themselves be ineligible formed earlier, typically outside of mar- for social services). So-called middle-class riage, and they are less stable, with less families may well feel the effects of an

144 (2) Spring 2015 35 The Future aging society most acutely. It seems likely What will the health and economic sta- of Intergen- that more of these families will adopt the tus of future Americans be? How will the erational Relations coresidence strategy of low-income house- circumstances of the elderly be affected by in Aging holds in an effort to pool their resources. trends in health and disease and changes in Societies It is impossible to know how economic family structure, educational attainment, growth and wage trajectories for various and public and family support systems? job sectors will mitigate the pressures on The answers to these questions de pend families in different countries twenty or on a range of alternative future sce nar ios, thirty years out; however, the prospects do each with a different set of costs and ben- not look especially bright, given that many e½ts. Economic and demographic model - advanced economies are struggling to ing is a useful tool that provides a simpli - maintain the robust economic growth ½ed representation of our future reality. necessary for a good labor market of young A team of researchers from the Mac - and middle-aged adults that are capable Arthur Network in collaboration with the of helping to support the oldest generation. Schaeffer Center for Health Policy and Eco - Perhaps older persons–who are in better nomics at the University of Southern Cal- health than previous generations–will be ifornia has been building such a model: more amenable to remaining in the labor the Future Americans Model (fam). fam force through their late sixties and early is a simulation model that uses data on seventies, both to improve their own cir- groups of Americans over age twenty-½ve cumstances and to contribute resources to from the Panel Study of Income Dynamics younger generations. and the Health and Retirement Study in or- This much we do know: large parts of der to understand the short- and long-term the future elderly are likely to be less af - implications of changes across cohorts on fluent than current cohorts of individuals well-being from birth through death. This over sixty-½ve. And, in all likelihood, many model expands on the technical infrastruc- of their children will be less well off in their ture and expertise the researchers devel- middle years than their parents were at the oped while building and using a similar same age. Young adults today are as well- model, the Future Elderly Model (fem). educated (or better-educated) than their Researchers from the Network are also parents were, but they are entering adult- using the fem to quantify the value of cur - hood with more debt and more modest rent and future intergenerational support labor-market prospects. and resource-exchange for Americans aged sixty-½ve and older. They are simulating These speculations about the future of the effects of changes in pub lic assistance intergenerational relations are informed (via Social Security and Medicare) on these by current research, but there are still many exchanges of money and time across fam - unknowns. The demographic growth of ilial generations. Looking forward, the Net- the elderly population, followed by a far work researchers’ work will deepen our un - smaller cohort of younger and middle-aged derstanding of fu ture intergenerational adults, is ½xed, but we do not know exactly exchange and support by anticipating fu - how these cohorts will look in another ture changes to the family. twenty or thirty years as the baby-boom Part of the challenge of building the Fu - cohort ages. They will certainly be more ture Americans Model is taking into ac- ethnically diverse, but whether current count the ways in which new family pat- trends in economic inequality and family terns will emerge for each age stratum in change will persist is less clear. the next twenty years. But predicting fam -

36 Dædalus, the Journal ofthe American Academy of Arts & Sciences ily forms is fraught with challenges. For to meet them. How do parents respond? Frank F. this reason, the Network has given a grant Do they call on additional relatives such as Furstenberg, Caroline Sten to a second team of researchers (headed by grandchildren or siblings to help out? Do Hartnett, Martin Chris Sepaki and including Andrew Cher - they look to public services? Or does the Kohli & Julie M. lin, Emily Agee, and Douglas Wolfe) to demand simply go unmet? The Network Zissimopoulos build a framework for exploring how fam- will compare the patterns in these already ily forms may evolve over the decades. Such “older” societies with the circumstances changes may also vary over socioeconomic of midlife parents in the United States. strata (as well as family forms), affecting To supplement this survey, researchers the potential flow of family assistance in supported by the Network will be conduct- the near future. The team will begin with ing in-depth interviews of family mem bers different assumptions and an alternate ap- in a three-generation qualitative study car- proach to that of the fam team, using cur- ried out by human development researcher rent information about the type and pace Karen Fingerman and her colleagues.23 of changes in marriage, cohabitation, and The aim of this research is to examine in fertility within and outside of wedlock. depth how exchanges of support are initi- Both teams are exchanging information so ated and responded to in daily life when the that they can determine whether and how oldest generation be gins to need more as - different assumptions about family change sistance. In research on intergenerational and the growth of inequality may affect exchange, very little attention has been public and private demands for health care paid to the needs families are unable to and economic support. In the next year, meet and what might remedy this short- they plan to publish the results from these fall. This qualitative research may help us simulations. understand how and when gaps occur be - This information will, in turn, give the tween the needs of the elderly and the sup- ½rst team (which is modeling future pat- ply of assistance. terns of costs and care for the elderly) a more speci½c and nuanced understanding This essay has sketched a picture of a de - of how demands on midlife parents and mographic future that may pose consider - their children may overlap with the needs able challenges for family systems in West - of the baby-boom generation as they age ern nations that during the twentieth cen - and require greater assistance from their tury worked to strike a balance between families. public and private support. That balance A third project, supported jointly by the is likely to change, and the economic and Network and the Pew Foundation, involves health costs of supporting the elderly gen- data collection on a largely unexplored is - eration while investing in younger genera- sue: how families in societies with more tions will become formidable over the next advanced population aging–such as Ger- thirty years. We have observed that this task many and Italy–are managing to deal with is likely to become more complex with the the simultaneous demands of caring for aging of the baby-boom cohort, changes their children and their elderly parents. The in family systems, and the growing strati - emphasis is on measuring unmet or par- ½cation of social classes and ethnic groups. tially met needs of individuals in the fami- Despite this essay’s overarching focus ly unit. As we discussed earlier, we an tici - on the problems, there are some silver lin - pate that, in an aging society, midlife par- ings to the current demographic situation. ents in particular will face demands for The relative paucity of children (G3s) com - assistance that are beyond their capacity pared to the elderly (G1s) could mean that

144 (2) Spring 2015 37 The Future although seniors will be less ½nancially nario of declining support for the elderly of Intergen- well-off on average, there will be a larger that we have projected. Nonetheless, the erational Relations cohort of them to provide assistance to American public, unlike its European coun- in Aging their children and grandchildren. The ratio terparts, appears to be relatively unaware Societies of seniors to grandchildren is likely to soar of the looming problems facing it as the in the next ½fteen to twenty years, suggest - baby-boom cohort reaches an ad vanced ing the possibility of a greater flow of par- age.24 The pressures on an already over- ent and grandparent assistance before the burdened and rapidly changing family sys - older generation requires assistance, by tem suggest that we could begin to see a which point grandchildren may have be- diminution (if not a full reversal) of the come a more potent source of assistance to downward flows of assistance from elderly the elderly. In fact, we know very little parents and relatives to younger genera- about how flows of assistance to the young tions. Were this to happen, it could restrict are reciprocated over longer stretches of social mobility and become an added time. source of inequality in the United States, We do not dismiss the possibility that sig- which is already among the most socio - ni½cant social, economic, and demograph- economically unequal of all econom ically ic changes could alter the rather bleak sce - developed nations.

endnotes * Contributor Biographies: FRANK F. FURSTENBERG, a Fellow of the American Academy since 1994, is the Zellerbach Family Professor of Sociology Emeritus at the University of Pennsyl- vania. He has authored or edited numerous books, including Destinies of the Disadvantaged: Teenage Childbearing and Public Policy (2007); Managing to Make It: Urban Families in High-Risk Neighborhoods (with Thomas D. Cook, Jacquelynne Eccles, and Glen H. Elder, Jr., 1999); and The New American Grandparent: A Place in the Family, a Life Apart (with Andrew J. Cherlin, 1986). His most recent book is Behind the Academic Curtain: How to Find Success and Happiness with a Ph.D. (2013). CAROLINE STEN HARTNETT is an Assistant Professor of Sociology at the University of South Carolina. Her research has been published in such journals as Population Studies, Demo- graphic Research, and Sociological Quarterly. MARTIN KOHLI is Distinguished Bremen Professor at the Bremen International Graduate School of Social Sciences (Germany) and Professor of Sociology, Emeritus, at the European University Institute, Florence (Italy). His work has been published in such journals as European Sociological Review, Demographic Research, Research in Human Development, and Journal of Euro- pean Social Policy, as well as in the recent editions of the Handbook of Aging and the Social Sci- ences (ed. Robert H. Binstock and Lisa K. George). JULIE M. ZISSIMOPOULOS is an Assistant Professor at the Sol Price School of Public Policy and the Associate Director of the Schaeffer Center at the University of Southern California. Her research has recently appeared in such journals as The Journal of Economic Perspectives, Jour- nal of Health Economics, Journal of Human Resources, and Demography. 1 John C. Caldwell, “Toward a Restatement of Demographic Transition Theory,” Population and Development Review 2 (3–4) (1976): 321–366. 2 United Nations Department of Economic and Social Affairs, World Youth Report 2003: The Global Situation of Young People (New York: United Nations, 2004).

38 Dædalus, the Journal ofthe American Academy of Arts & Sciences 3 Martin Kohli, Marco Albertini, and Harald Künemund, “Linkages among Adult Family Gen- Frank F. erations: Evidence from Comparative Survey Research,” in Family, Kinship and State in Con- Furstenberg, temporary Europe (Vol. 3: Perspectives on Theory and Policy), ed. Patrick Heady and Martin Kohli Caroline Sten Hartnett, Martin (Frankfurt: Campus Verlag, 2010), 225–248. Kohli & Julie M. 4 Marco Albertini, Martin Kohli, and Claudia Vogel, “Intergenerational Transfers of Time and Zissimopoulos Money in European Families: Common Patterns–Different Regimes?” Journal of European So- cial Policy 17 (4) (2007): 319–334. 5 Dennis P. Hogan, David J. Eggebeen, and Clifford C. Clogg, “The Structure of Intergenera- tional Exchanges in American Families,” The American Journal of Sociology 98 (6) (1993): 1428 –1458; and Debra Umberson, “Relationships between Adult Children and their Parents: Psychological Consequences for Both Generations,” Journal of Marriage and the Family 54 (3) (1992): 664–674. 6 Yoshinori Kamo, “Racial and Ethnic Differences in Extended Family Households,” Sociological Perspectives 43 (2) (2000): 211–229. 7 Pew Research Center, “The Decline of Marriage and Rise of New Families” (Washington, D.C.: Pew Research Center, 2010), http://www.pewsocialtrends.org/2010/11/18/the-decline-of -marriage-and-rise-of-new-families/. 8 Daphna Gans, Ariela Lowenstein, Ruth Katz, and Julie Zissimopoulos, “Is There a Trade-Off Between Caring for Children and Caring for Parents? The Nexus of Informal and Formal Care for Older Adults,” Journal of Comparative Family Studies 44 (4) (2013); and Judith Seltzer and Suzanne M. Bianchi, “Demographic Change and Parent-Child Relationships in Adult- hood,” Annual Review of Sociology 39 (2013): 275–290. 9 Emily Grundy and John C. Henretta, “Between Elderly Parents and Adult Children: A New Look at the Intergenerational Care Provided by the ‘Sandwich Generation,’” Ageing and Society 26 (5) (2006): 707–722. 10 Emily Wiemers and Suzanne Bianchi, “Sandwiched between Aging Parents and Boomerang Kids in Two Cohorts of American Women,” University of Massachusetts Boston Working Paper 2014-06 (Boston: University of Massachusetts Boston Department of Economics, 2014). 11 Joyce A. Martin, Brady E. Hamilton, and Michelle Osterman, Three Decades of Twin Births in the United States, 1980–2009 (Hyattsville, Md.: National Center for Health Statistics, Centers for Disease Control and Prevention, 2012); Joyce A. Martin, Brady E. Hamilton, Michelle J. K. Osterman, Sally C. Curtin, and T. J. Matthews, Births: Final Data for 2012 (Hyattsville, Md.: National Center for Health Statistics, Centers for Disease Control and Prevention, 2013); and T. J. Mathews and Brady E. Hamilton, Delayed Childbearing: More Women Are Having Their First Child Later in Life (Hyattsville, Md.: National Center for Health Statistics, Centers for Disease Control and Prevention, 2009). 12 Patrick D. Wightman, Megan Patrick, Robert Schoeni, and John Schulenberg, Historical Trends in Parental Financial Support of Young Adults (Ann Arbor: University of Michigan, 2013). 13 Andrew J. Cherlin, “The Deinstitutionalization of American Marriage,” Journal of Marriage and Family 66 (4) (2004): 848–861; and Patrick Heuveline and Jeffrey M. Timberlake, “The Role of Cohabitation in Family Formation: The United States in Comparative Perspective,” Journal of Marriage and Family 66 (5) (2004): 1214–1230. 14 Pew Research Center, “The Decline of Marriage and Rise of New Families.” 15 Teresa M. Cooney and Peter Uhlenberg, “The Role of Divorce in Men’s Relations with Their Adult Children after Mid-Life,” Journal of Marriage and Family 52 (3) (1990): 677–688; Matthijs Kalmijn, “Gender Differences in the Effects of Divorce, Widowhood and Remarriage on Inter - generational Support: Does Marriage Protect Fathers?” Social Forces 85 (3) (2007): 1079– 1104; and Liliana E. Pezzin and Barbara Steinberg Schone, “Parental Marital Disruption and Intergenerational Transfers: An Analysis of Lone Elderly Parents and Their Children,” De - mography 36 (3) (1999): 287–297.

144 (2) Spring 2015 39 The Future 16 Lindsay Monte and Renee Ellis, Fertility of Women in the United States: 2012 (Washington, D.C.: of Intergen- U.S. Census Bureau, 2014); and Pew Research Center, “The Decline of Marriage and Rise of erational New Families.” Relations in Aging 17 Ingrid Connidis, “Contact between Siblings in Later Life,” Canadian Journal of Sociology/ Societies Cahiers canadiens de sociologie 14 (1989): 429–442; Pearl A. Dykstra, “Childless Old Age,” in International Handbook of Population Aging, ed. Peter Uhlenberg (New York: Springer, 2009), 671–690; Colleen Leahy Johnson and Donald J. Catalano, “Childless Elderly and Their Family Supports,” The Gerontologist 21 (6) (1981): 610–618; Martin Kohli and Marco Albertini, eds., Minimal Families: Childlessness and Intergenerational Transfers (Special Issue of Ageing & Society) 29 (8) (2009): 1171–1274; and Robert L. Rubinstein et al., “Key Relationships of Never Married, Childless Older Women: A Cultural Analysis,” The Journals of Gerontology: Psychological Sciences & Social Sciences 46 (5) (1991): S270–S277. 18 Sara McLanahan, “Diverging Destinies: How Children Are Faring under the Second Demo- graphic Transition,” Demography 41 (4) (2004): 607–627; and Sara McLanahan and Christine Percheski, “Family Structure and the Reproduction of Inequalities,” Annual Review of Sociology 34 (2008). 19 S. Philip Morgan and Ronald R. Rindfuss, “Reexamining the Link of Early Childbearing to Marriage and to Subsequent Fertility,” Demography 36 (1) (1999): 59–75; R. Kelly Raley and Larry Bumpass, “The Topography of the Divorce Plateau: Levels and Trends in Union Stability in the United States after 1980,” Demographic Research 8 (8) (2003): 245–260; Ronald R. Rind- fuss, S. Philip Morgan, and Kate Offutt, “Education and the Changing Age Pattern of American Fertility: 1963–1989,” Demography 33 (3) (1996): 277–290; and Judith A. Seltzer, “Relationships between Fathers and Children Who Live Apart: The Father’s Role after Separation,” Journal of Marriage and Family 53 (1) (1991): 79–101. 20Frank F. Furstenberg, “Fifty Years of Family Change: From Consensus to Complexity,” The ANNALS of the American Academy of Political and Social Science 654 (2014): 12–30. 21 Richard Fry and Jeffrey S. Passel, “In Post-Recession Era, Young Adults Drive Continuity Rise in Multi-Generational Living,” Pew Research Center Social and Demographic Trends Project (Washington, D.C.: Pew Research Center, 2014). 22 Robert F. Schoeni and Karen E. Ross, “Material Assistance from Families during the Transition to Adulthood,” in On the Frontier of Adulthood: Theory, Research, and Public Policy, ed. Richard A. Settersten, Jr., Frank F. Furstenberg, and Rubén G. Rumbaut (Chicago: University of Chicago Press, 2005), 396–417. 23 K. L. Fingerman et al., “Who Gets What and Why: Help Middle-Aged Adults Provide to Par- ents and Grown Children,” The Journals of Gerontology: Psychological Sciences & Social Sciences 66 (2011): 87–98. 24 Pew Research Center, “Attitudes about Aging: A Global Perspective” (Washington, D.C.: Pew Research Center, 2014), http://www.pewglobal.org/2014/01/30/attitudes-about-aging-a-global -perspective/.

40 Dædalus, the Journal ofthe American Academy of Arts & Sciences Labor-Force Participation, Policies & Practices in an Aging America: Adaptation Essential for a Healthy & Resilient Population

Lisa F. Berkman, Axel Boersch-Supan & Mauricio Avendano

Abstract: Population aging in the United States poses challenges to societal institutions while simultaneously creating opportunities to build a more resilient, successful, and cohesive society. Work organization and labor-force participation are central to both the opportunities and challenges posed by our aging society. We argue that expectations about old age have not suf½ciently adapted to the reality of aging today. Our institutions need more adaptation in order to successfully face the consequences of demographic change. Although this adaptation needs to focus especially on work patterns among the “younger elderly,” our society has to change its general attitudes toward work organization and labor-force participation, which will have implications for education and health care. We also show that work’s bene½cial effects on well-being in older ages are often neglected, while the idea that older workers displace younger workers is a miscon- ception emerging from the “lump-of-labor” fallacy. We conclude, therefore, that working at older ages can lead to better quality of life for older people and to a more productive and resilient society overall.

Population aging in the United States, as in most countries around the world, poses many challenges LISA F. BERKMAN is the Director to our major private, public, and societal institutions. of the Harvard Center for Popula - At the same time, however, it creates great oppor- tion and Development Studies. tunities for building a more resilient, successful, and AXEL BOERSCH-SUPAN is the Di - cohesive society. For several reasons, work organiza - rector of the Munich Center for the tion and labor-force participation are central to both Economics of Aging at the Max opportunities and challenges posed by an aging pop - Planck Institute for Social Law and ulation. First, for all but a very few, working remains Social Policy. the main source of income for consumption and for MAURICIO AVENDANO is Princi- savings. Work also has an anchoring function in pal Research Fellow and Deputy Di- society, bringing multiple bene½ts to our physical, rector of lse Health at the London social, and emotional well-being. Given the continu- School of Economics and Political ing increase in life expectancy in aging societies, it is Science. more necessary than ever before to revisit the role (*See endnotes for complete contributor of work in older ages and the opportunities that biographies.) longer working lives can bring to aging societies. We

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00329 41 Labor-Force argue in this essay that individual expec- of these issues, beginning with the salient Participation, tations about old age have not suf½ciently points about Social Security (a more de - Policies & Practices in adjusted to the new reality of an aging soci - tailed discussion of which may be found an Aging ety. Reflecting this, our private, public, and in S. Jay Olshansky, Dana P. Goldman, and America societal institutions suffer from the same John W. Rowe’s essay in this volume.) disconnect and now need to adapt consid- At the time Social Security was estab- erably to face the challenges and embrace lished in 1935, sixty-½ve was commonly the the opportunities of demographic change. age U.S. citizens received other govern- Although this adaptation must focus espe - ment bene½ts. (It is important to recall that cially on work patterns among the “youn - life expectancy for American men from ger elderly,” our entire society has to adapt 1935 to 1940 was about sixty years.) Table 1 its attitudes toward work organization and shows the dramatic changes in life expec - labor-force participation, and in the pro - tancy that have taken place since then– cess rethink its education and health care changes that have been reported in many policies and expectations. articles and studies, but that appear not to In this essay, we will describe some of the have entered our collective wealth of self- often understated bene½ts that working evident facts on which individual actions and remaining active may have for health and general policy are based. and well-being in older populations. Fur- thermore, it is often claimed that an older In the mid-twentieth century, just over workforce has negative implications for half of all Americans who reached the age general productivity and displaces younger of twenty-one could expect to reach sixty- workers from positions they would other- ½ve. Many workers paying into Social Se- wise occupy. We will show that this is a mis- curity would never live long enough to re- conception emerging from the “lump-of- ceive bene½ts, especially African Ameri- labor” fallacy: the idea that the amount cans, whose life expectancy was just over of work available to laborers is ½xed. We ½fty in 1935. Today, however, men who re- maintain that it is not demographic tran- tire at age sixty-½ve can expect to live for sitions per se that will shape our future, an additional seventeen years; women live but instead how our institutions and poli - even longer.1Average life expectancies have cies respond and adapt to them. It is our risen so sharply since 1935 that it is no sur - choice. prise that Social Security–and our expectations about what old age looks like–have Individual expectations about old age been unable to keep up. tend to be formed by looking to history. Signi½cantly, there is also evidence for Our intuition about health and workabil- a substantial increase in healthy and dis- ity at older ages stems to a large extent ability-free life expectancy; that is, the from the experiences of our parents and num ber of years men and women can ex - grandparents. Our expectations about pect to live without major functional lim- what Social Security should deliver to us itations. Despite some controversy, the stem from what Social Security delivered gen eral trend appears clear: over the past to previous generations. We have a hard decades, as life expectancy has improved time imagining, however, what our lives in the United States, so too has healthy or will look like in twenty-½ve or ½fty years: disability-free life expectancy. When life how long we will live, how healthy we will expectancy improves and morbidity and be by then, and under how much pressure disability decease, this leads to “compres- Social Security will be. We will address all sion of morbidity,” or a longer part of life

42 Dædalus, the Journal ofthe American Academy of Arts & Sciences Table 1 Lisa F. Life Expectancy for Social Security Berkman, Axel Boersch- Year Cohort Percentage of Population Average Remaining Life Expectancy Supan & Turned 65 Surviving from Age 21 to Age 65 for Those Surviving to Age 65 Mauricio Avendano Male Female Male Female 1940 53.9 60.6 12.7 14.7 1950 56.2 65.5 13.1 16.2 1960 60.1 71.3 13.2 17.4 1970 63.7 76.9 13.8 18.6 1980 67.8 80.9 14.6 19.1 1990 72.3 83.6 15.3 19.6

Source: Social Security Administration, “Life Expectancy for Social Security,” http://www.ssa.gov/history/ lifeexpect.html. spent in good health and a shorter number narios of population trends in disability– of years lived in ill health. Most studies that those with a high school diploma have show that a signi½cant compression of generally a high and consistent potential mor bidity has occurred over the last de - to work productively between the ages of cades in the United States, making it clear ½fty-½ve and seventy-four. The estimates that many–although not all–Americans suggest that under most scenarios, about are able to work until after current retire- 70 percent of those with at least a high ment ages or eligibility for Social Security.2 school diploma will be able to work in Yet some evidence suggests that com- this age range. These ½ndings are generally pression of morbidity has stagnated in re - in line with studies done by economist cent years. Even more concerning are cur - David M. Cutler and others, who assume rent hints that the middle-aged and the no huge upward turns in disability preva- “young old” are showing increases in mor- lence, even with modest or no improve- bidity and functional limitations not evi- ments in active life expectancy.4 dent in earlier cohorts that grew old be - But the same is not true for those with tween 1990 and the present. A closer look low levels of education. Rehkopf’s team at the evidence shows that good health at sug gests that if less favorable trends con- older ages is strongly predicted by educa- tinue for men and women without a high tional attainment and other indicators of school diploma, this group could have de- socioeconomic status. In a recent study, pressed levels of ability to work at older public health researcher David Rehkopf ages.5 These estimates vary much more and colleagues projected the future em- depending on different assumptions about ployment potential for the near-elderly trends, but they range from just over 0 per - workforce based on demographic trends cent to 60 percent in the most optimistic and trends in mobility and functional sta- circumstances. Of course, new medical tus from the 1982–2004 National Long tech nologies could alter these trends by Term Care Study and the 1992–2010 providing more support to those with Health and Retirement Study.3 Rehkopf’s limitations. Furthermore, patterns of im - projections through 2050 focus speci½c at - mi gration bringing in more or less educated tention on educational differences. His young and middle-aged adults could also group concludes–based on multiple sce- change these estimates. But, assuming that

144 (2) Spring 2015 43 Labor-Force less dramatic changes take place, we can terns and demographic reality. In response, Participation, anticipate seeing a divided workforce: one France and Italy have joined the growing Policies & Practices in group of men and women with a high po - number of European countries adapting an Aging tential to work in their sixties and seven- their statutory retirement ages, although America ties, and another whose more substantial only for future cohorts and with many disabilities limit their potential for long- exceptions. Germany, for instance, has con - term labor-force participation. This divi- fronted this challenge through a variety of sion in potential future ability to work policy adjustments: First, it introduced makes it even more dif½cult to form ex - penalties for early retirement that were pectations about the type of support an gradually phased in between 2000 and aging society will need. Political discus- 2010. Second, it abolished some special sions tend to focus on one or the other tracks for early retirement, including the group, stressing either the disadvantaged differential treatment of men and women group’s need to retire early or the healthier (who previously could retire earlier despite group’s need (or at least ability) to work their longer life expectancy). Third, the longer. Adapting work organization and statutory retirement age began gradually labor-force participation to an aging shifting in 2013 by about a month per year world, however, requires recognition of to reach a target of age sixty-seven by 2029 U.S. society’s deep split between the many –roughly in line with the expected in - for whom healthy life expectancy has dra- crease in the length of life. These adapta- matically increased, and those who have tions have met much resistance. Accom- not enjoyed such improvements in health. modating pop ular opinion, the German gov ernment reintroduced a pathway to The dramatic demographic changes de- early retire ment at age sixty-three. France picted in Table 1 imply that most men, and and Italy have seen similar policy reversals, even more women, will now survive early backtracking from modest increases in stat - childhood and mid-adulthood to reach utory retirement ages to early retirement old er ages, meaning that larger and larger op tions. Pub lic anxiety about increased re - numbers of older men and women will tirement age clearly shows the need to over - reach eligibility for Social Security. When come popular misconceptions about actu- most retirement and other age-related pol- al and fu ture life expectancy and health in icies were established, they made sense in older ages. terms of the current life expectancy. Today The United States has done part of its they do not. These demographic changes homework and is ahead of Europe in its point to the necessity of institutional adap- preparations. The Social Security Amend- tation. ments of 1983 legislated a gradual shift in Among industrialized nations, the Unit- the eligibility age for normal Social Secu- ed States does not have the largest dispar- rity bene½ts. It will not be suf½cient, how - ity between life expectancy and retirement ever, to bring the U.S. Social Security sys- age. France and Italy, for example, set the tem back into actu arial balance. Thus, a age of early retirement earlier than the steeper path to changing the average work- United States’ (often between ½fty-½ve and er’s eligibility ages is still necessary. sixty), even though they enjoy greater life In addition, given the reality of a divide expectancy than the United States does. in the health of America’s older popula- Thus, the time from retirement to death tion, it is critical to create differentiated is decades long in France and Italy, creat- paths to retirement and labor-force exits ing a large imbalance between work pat- depending upon health (which in turn of-

44 Dædalus, the Journal ofthe American Academy of Arts & Sciences ten depends on economic and social exper - adopt a life-course approach that focuses Lisa F. iences earlier in life). This life-course fork on the social experiences and behaviors of Berkman, Axel Boersch- in long-term trajectories suggests that op - men and women across all ages. For ex - Supan & tions for early retirement based on work ample, the men and women who will be in Mauricio disability are critical to maintaining intra- their sixties and seventies in 2030 to 2050 Avendano generational equity between those who can are in early to mid-adulthood now. Their work longer and those who cannot. A one- current health and social conditions are size-½ts-all Social Security reform will therefore shaping their capacity and the op- not do; only a differentiated approach will portunities for employment options they avoid the political backlashes observed in will have at age sixty, seventy, or eighty. Europe. This may mean the implementa- The point is even stronger for education tion of both a general retirement age that because of the many implications a poor is indexed in some way to life expectancy education has on opportunities over the and an early-retirement option based on entire course of life. We have no time to the ability to work. For older workers in lose if we hope to change trajectories of poor health, it is obviously better for their work and retirement. health and well-being not to have to work. Debates about our aging society too This may mean that certain groups within quickly end up foundering on the issue of the population–such as the less educated Social Security reform and the well-being and those with very physically demanding of the elderly, while tending to ignore the jobs–may need the option to take an early necessity of creating better starting posi- path to retirement. Current Social Security tions for the young. Macroeconomic anal - policies enabling people to start collecting yses show that education and health care bene½ts around age sixty-two may be a reform have more leverage in shaping our good model for workers in this group. society’s ability to create new jobs, foster Those with better functioning or with lim- better working conditions, and encourage itations that do not interfere with their per- labor-force participation (and thus, indi- formance could have longer working lives rectly, improve the ½nancial position of and continue to make substantial contri- Social Security) than an adapted retire- butions to productivity in U.S. society. ment age and disability insurance reform.6 Other studies suggest that recessions have Adapting the retirement age to a longer differential impacts on health in the long life has many implications. We note, as run, with disadvantages accruing to par- others in this volume have, that these do ticularly vulnerable age groups.7 not exclusively affect the elderly. First, the ability of future generations to work longer In any argument about retirement, it is hinges on their education and health important to discuss the implications of throughout life. Additionally, since the working at older ages. In general, being United States will have a larger number employed is positively associated with of retirees in the future than ever before, health. There are many reasons for this re- the younger and middle-aged will have to lationship, the most obvious being health be better-educated and in better health selection; that is, healthy people are more than they are today in order to carry this likely to be able to work. However, there is additional actuarial burden (regardless of growing evidence that employment itself how many older people keep working). actively yields both physical and mental To optimize our chances of providing a health bene½ts. Here, we will explore ½nd- healthier workforce at older ages, we must ings suggesting that employment may im -

144 (2) Spring 2015 45 Labor-Force prove health and well-being by increasing cy of retirement bene½ts, as well as indi- Participation, social engagement; developing and main- vidual factors such as occupation, socio- Policies & Practices in taining intellectual and interpersonal skills; economic status, and marital status. Retire- an Aging and, importantly, continuing to earn and ment may also have different effects on America delaying the use of savings, pensions, and physical and mental health, requiring the other bene½ts. This phenomenon has ob- need for further differentiation in the vious implications for retirement, since methodology of studies on retirement. negative effects of retirement are often ig- The literature on this question has fo - nored in the cost-bene½t analysis that is cused on understanding how reforms on done when a state attempts to determine the age of retirement–namely, the age of the optimal retirement age. compulsory retirement or the minimum Retirement introduces large changes to age of retirement–might impact health. an individual’s life. While there is little These laws have a strong effect on retire- doubt that poorer health is associated with ment decisions: while a substantial pro- early retirement, studies on the health im - portion of workers retire before the stat - pact of retirement have so far reached no utory age of retirement, a higher statutory consensus on whether retirement pro- retirement age encourages individuals to motes or harms health. Differentiation is work longer. Here we focus on a small (but essential in conducting these studies: growing) set of studies that have attempted phys ically and psychologically strenuous to establish causality between the statu- work conditions are unlikely to be good tory retirement age and retirement’s ef- for a person’s health, while working in a fects on health. We will examine two rewarding and healthy environment may pieces of evidence that are important in be better for mental and physical health this respect: First, we discuss some of the than leaving the workforce. Many of the studies using longitudinal data to assess existing studies have faced methodologi- how retirement influences health, taking cal problems: they do not distinguish the into account the complex set of factors effects of aging from those of retirement that lead individuals into retiring early. and they often do not distinguish the ef - Sec ond, we discuss the evidence of the fects of retirement from those of previous health impact of retirement age–related life experiences or conditions that them- policy reforms. selves influence retirement decisions. Many of the descriptive studies lack an Longitudinal studies follow the health ad equate control or comparison group. of workers during the years prior, during, This is important because the decision to and after retirement and compare it to the retire is not random: there is a self-selection health trajectories of workers who contin- issue. For instance, those who are ill are ue to work. Using data from employees more likely to retire early. Those with great from the French companies Électricité de wealth may also retire early. However, ill- France and Gaz de France (edf-gdf; also ness may be caused by environmental con- called the gazel cohort), epidemiologist ditions and wealth by enhanced educa- and gerontologist Hugo Westerlund and tional opportunities. Thus, in both cases, colleagues found that between the year retirement is not causally linked to health before and the year after retirement, the or wealth but to previous life experiences prevalence of poor self-rated health fell and conditions. In addition, the effects of from 19 percent to 14 percent.8These health retirement on health may depend on many improvements were stronger for workers contextual factors, including the adequa- with a poor work environment before re-

46 Dædalus, the Journal ofthe American Academy of Arts & Sciences tirement. Using data from the Whitehall II ½ts. Based on these variations, economists Lisa F. study of British civil servants, epidemiol- Norma B. Coe and Gema Za mar ro found Berkman, Axel Boersch- ogist Gill Mein and colleagues found that that retirement leads to a short-term de- Supan & although mental health improved after crease in the probability of reporting poor Mauricio retirement, physical functioning did not health, and a long-lasting improvement in Avendano appear to change.9 Mental health improve- the overall health index. 12 ments, however, were con½ned to high- grade employees. In a reexamination of the It is nonetheless important to distinguish data, epidemiologist and social scientist general health from cognitive abilities, Markus Jokela and colleagues found that which appear to bene½t from working at compulsory retirement at age sixty and older ages. The impact of retirement on early voluntary retirement were associated cognitive function is of particular inter- with improvements in mental health and est in view of the “use it or lose it” hypo - physical functioning.10 In contrast, retire- thesis, which suggests that age-related cog- ment due to ill health was associated with nitive decline can be lessened through en - poorer mental health and physical func- gagement in cognitively demanding ac - tioning. Their ½ndings highlight the im - tivities.13 This suggests that individuals portant role of health-related selection as whose work is cognitively demanding may a potential explanation of the negative asso- bene½t from later retirement. Using data ciation between retirement and health. from the Whitehall II study, B. A. Roberts Many of the studies referenced above, and colleagues found that workers who which are con½ned to European popula- retired experienced smaller improvements tions, cast doubt on the notion that retire - in mean cognitive test scores than contin- ment is bad for health overall: the pre- uously employed workers, although these vailing ½nding appears to be that in the differences were not signi½cant for most short term, retirement is associated with cognitive test scores.14 A seminal paper an improvement in mental health and lit- by psychologist Stéphane Adam and col- tle or no change (but no clear evidence of leagues, based on the Study of Health, harm) to physical health; though there is Aging, and Retirement in Europe (share), no doubt that the effect of retirement de - found that cognition–measured mainly pends on the nature of the worker’s occu - by memory abilities such as delayed word pation and health prior to retirement. recall–declined during retirement.15 This More recently, studies have used differ- ½nding has given rise to an entire litera- ences across cohorts in eligibility for retire- ture. Economists Susann Rohwedder and ment bene½ts (based on legislation on stat- Robert J. Willis compared studies of re - utory retirement or pensionable ages) to tirement from Europe, the United States, isolate the effects of retirement on health. and Britain and found that early retirement Economist Kerwin Ko½ Charles has used has a negative effect on cognitive ability.16 policy variation in mandatory retirement Their ½ndings, however, are not corrobo- and Social Security bene½ts that influence rated by other studies. Based on data from retirement incentives by age and cohort in the U.S. Health and Retirement Study the United States to examine the impact of (hrs), Coe and colleagues ex amined em - retirement on depression.11 He found that ployers that offered early retirement win- retirement leads to better mental health dows and found that time in retirement and well-being. Other studies ex ploit var - was unrelated to cognitive function among iations across countries in the age of eligi- white-collar workers, but may have a pos - bility for early and full retirement bene- itive effect on cognitive function among

144 (2) Spring 2015 47 Labor-Force blue-collar workers.17 More recent studies often said to have negative side effects for Participation, based on European data, however, show the economy as a whole. For many years, Policies & Practices in that early retirees suffer from faster cogni- common sense suggested that the number an Aging tive decline than later retirees.18 Based on of jobs in the economy is ½nite, and that a America the hrs, sociologist Esteban Calvo and new population entering the labor force colleagues found that the effect of retire- would therefore push other workers out. ment depends on the timing: those retir- This so-called lump-of-labor fallacy has ing before age sixty-two seem to fare worse been invoked at moments in history when than those who continue to work; yet, re - women’s labor-force participation in - tirement at age sixty-two or older is not creased, because it was thought that they associated with worse physical and mental would take “good jobs” away from men. health.19 Again, these studies have gener- Immigrants to the United States continue ally not found any negative effects of retire - to be accused of stealing jobs from other, ment on physical or mental health. Other native lower-wage workers. Likewise, studies have also found that retirement has many older people who wish to continue no impact on mortality.20 working today are accused of taking jobs The studies discussed above reflect the from younger workers, creating intergen- variety of approaches to determining the erational conflict. The lump-of-labor fal - health effects of retirement. What do these lacy is one of the most damaging myths in results tell us about the potential impact economics.21 It is deeply rooted in the be - of recently enacted policies to increase re- lief that the economy resembles a small tirement age for future generations across enterprise with a small, ½xed number of many countries? On the one hand, there clients and a ½xed demand for its product. seems to be little evidence that retirement Such an enterprise has a set amount of harms physical health or increases the risk output based on demand, and therefore of dying. Although some studies do sug- can only use a certain amount of labor. gest that retirement may be bene½cial to This is a poor analogue to a suf½ciently mental health, distinguishing between dif - large and complex economy. This is shown ferent cohorts is again paramount: sever- most clearly in the United States, where al studies suggest that the mental health the sharp increase in female labor-force consequences of retirement depend on the participation not only did not cause mass working environment and type of job the un employment for men, but actually cor- retiree had. While retirement does ap pear related with a rise in male employment to bene½t the mental health of many work - rates. More speci½cally, recent ½ndings ing in strenuous conditions and perform- from cross-national comparisons show ing manual labor, this is less clear for that higher employment of older individ- work ers in white-collar positions and with uals is actually positively correlated with healthy work environments. Finally, the higher em ployment of the young; that is, more recent evidence tends to ½nd some countries with a high prevalence of early support for the hypothesis that retiring re tire ment tend to have higher unemploy - later helps individuals maintain better cog- ment rates and lower employment of the nitive function. young.22 Figure 1 shows a correlation between What effects will delayed retirement early retirement and youth unemployment have on the greater population and the in oecd countries. These ½ndings may be young in particular? Higher rates of labor- challenged, however, as many confound- force participation in older individuals is ing factors operate at the same time in the

48 Dædalus, the Journal ofthe American Academy of Arts & Sciences Figure 1 Lisa F. Early Retirement and Unemployment in the oecd Berkman, Axel Boersch- Supan & 70.0 Mauricio Hungary Avendano Belgium 60.0 Austria Poland Italy France Greece 50.0 Czech Republic Spain

Netherlands United Kingdom 40.0 Finland Denmark Germany Estonia Japan United States 30.0 Switzerland Sweden 20.0

10.0

Percent of Elderly Out Labor Force 0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 Percent of Youth Unemployed

The R-squared of the correlation (a measure of statistical tightness) is 18 percent. This is not driven by the out- liers (Greece and Spain). If these crisis-affected economies are omitted, the positive correlation is actually stronger (with an R-squared of 21 percent). Source: Figure prepared by authors using calculations based on the 2012 Or - ganisation for Economic Co-operation and Development, Employment Outlook 2012 (Paris: oecd Publishing, 2012). aggregate data. Strong and isolated re forms duced in 1984 substantially increased the are more suitable for empirically identify- unemployment rate of those aged ½fty-½ve ing the effects of pension policies on labor- to ½fty-nine, since unemployment insur- market outcomes for the young. It is there- ance bene½ts were used as substitutes for fore instructive to examine the impact of early retirement pensions. Yet youth em- speci½c pension reforms on employment ployment did not rise in response. The rates at different ages. phasing-in of “actuarial” adjustments after Germany provides a particularly neat 1998 re versed the trend of early retirement. case, since strong and isolated reforms in Employment increased from 30 percent the years 1972, 1984, and 1998 dramatically to 40 percent in those aged sixty to sixty- changed retirement incentives.23 Figure 2 four. There is a very slight concurrent de - depicts the labor-force participation rates crease in employ ment of the young. for four age groups in Germany, and Figure The ½rst two cases are clear-cut: employ- 3 presents the corresponding unemployment of the young and the old moved in ment rates. These ½gures reveal three im - tandem. But the third case may appear to portant facts. First, the 1972 reform dra- contradict this relationship. Axel Boersch- matically reduced retirement age, labor- Supan and economist Reinhold Schnabel, force participation, and employment of however, have shown in their regression old er individuals. In spite of this, youth analysis of the third case that the slight de- employment did not increase. Second, the crease in employment of the young is in “bridge to retirement” legislation intro- fact a reflection of the business cycle and

144 (2) Spring 2015 49 Labor-Force Figure 2 Participation, Labor-Force Participation of Young and Elderly Males in Germany, 1960–2006 Policies & Practices in 100% an Aging America 90% men 25-54 80%

70% men 55-59 60%

50% men 20-24

40%

Employment Rate 30% men 60-64

20%

10%

0% 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Year

Each line represents a different age group. Source: German Mikrozensus, https://www.destatis.de/DE/Methoden/ SUFMikrozensus.html.

Figure 3 Labor-Force Participation of Young and Elderly Males in Germany, 1966–2006

all age 15-24 age 25-54 age 55-64

17% 16% 15% 14% 13% 12% 11% 10% 9% 8% 7% 6%

Unemployment Rate 5% 4% 3% 2% 1% 0% 1965 1975 1985 1995 2005 Year

Each line represents a different age group. Source: German Mikrozensus, https://www.destatis.de/DE/Methoden/ SUFMikrozensus.html.

50 Dædalus, the Journal ofthe American Academy of Arts & Sciences not a response to the introduction of ac - Security, possible cognitive bene½ts to Lisa F. tuarial adjustments.24 working at older ages, and the potential for Berkman, Axel Boersch- The German analysis is part of the work economic vitalization are some of the fac - Supan & by an international team that used pension- tors that support increasing the number of Mauricio design changes in eleven countries to iden- individuals who work past today’s statuto- Avendano tify how changes in the employment of ry retirement age. It would be naive to ex - older individuals has affected the employ - pect that this will occur only through Social ment of the young. The results vary consid- Security reform and legislation encourag - erably across speci½cations, but in these ing workers to work longer; we also need studies there are many more cases that re- structural policy changes that generate a fute the lump-of-labor hypothesis than healthier and more productive America. cases that support it. As economists Jona- These include policies that invest in human than Gruber and David Wise have written: capital throughout individuals’ lives, thus enabling them to work longer; such as poli - The overwhelming weight of the evidence, cies on early childhood, education, employ - as well as the evidence from each of the ment protection, work flexibility, income several different methods of estimation, is support, poverty reduction, and health contrary to the “boxed economy” proposi- care access.26 Most individuals should not tion. We ½nd no evidence that increasing experience deterioration of mental and the employment of older persons will phys ical health from working longer; reduce the employment opportunities of rather, the goal is to support healthy aging youth and no evidence that increasing the in such a way that working will be more employment of older persons will increase feasible and potentially flexible for older the unemployment of youth.25 cohorts. Only through policies that pro- Countries have large multifaceted econ - mote life trajectories leading to healthy omies that cannot be likened to small com- aging will we be able to create a workforce panies with ½xed, “boxed” labor needs. able to work longer, and only then will we National economies can grow, increasing be able to accrue the true societal bene½ts the demand for all goods and services and of So cial Security re form. therefore also the demand for labor. As How we adapt the major U.S. institutions with women and immigrants, the increas- related to work organization and labor- ing entry of older workers into the workforce participation will shape our future force contributes to a vital and productive as we move through this demographic economy. Moreover, unless a pension sys - transition. Although shaping public and tem is fully funded, there is a tax cost for pri vate policies is of paramount impor- retirement–whether early or not–that tance, this adaptation must take place on must be spread over the entire economy. all levels, such as with more informal work- This raises the total labor compensation place practices. These policies and prac- em ployers must pay for all workers, in - tices shape patterns of labor-force partic- cluding the young. The greater the num- ipation for older men and women and de- ber of older workers that leave the work- termine how they will pursue retire ment. force, therefore, the more likely it is that These policies must also take into account the employment prospects of the young that each older generation is a diverse set of will worsen. men and women with different life-course patterns of education, skills, family con- Increases in life expectancy and compres- stellations, and health conditions, and that sion of morbidity, funding de½cits in Social this in turn affects the em ployment op -

144 (2) Spring 2015 51 Labor-Force portunities they will have at older ages. Al - policy level and the other at the population Participation, so determining outcomes of labor policy level (related to the capacity of individu- Policies & Practices in changes are the social, economic, and als)–will determine whether we remain an Aging health capital of Americans in the labor a resilient and successful society as we ex- America force, currently and in the future. These perience the aging of our population. two sets of conditions–one at the labor-

endnotes * Contributor Biographies: LISA F. BERKMAN is the Thomas D. Cabot Professor of Public Policy and of Epidemiology at the Harvard School of Public Health and the Director of the Harvard Center for Population and Development Studies. She is the author or editor of numerous books and publications, including Social Epidemiology (second edition; edited with Ichiro Kawachi and Maria Glymour, 2014), Neighborhoods and Health (edited with Ichiro Kawachi, 2003), and Health and Ways of Living (with Lester Breslow, 1983). AXEL BOERSCH-SUPAN is the Director of the Munich Center for the Economics of Aging at the Max Planck Institute for Social Law and Social Policy. He has written numerous books and his articles have appeared in such journals as American Economic Review, Journal of Econometrics, Journal of Public Economics, and International Journal of Epidemiology. MAURICIO AVENDANO is Principal Research Fellow and Deputy Director of lse Health at the London School of Economics and Political Science, and Adjunct Associate Professor at Harvard University. His research has been published in such journals as American Journal of Epidemiology, Annual Review of Public Health, and Preventive Medicine. Authors’ Note: This work was supported by the MacArthur Foundation Research Network on an Aging Society and by a grant from the National Institute on Aging (5R01AG040248-03). Mauricio Avendano is also supported by the European Research Council (erc grant 263684). This paper includes sections adapted from Axel Boersch-Supan, “Myths, Scienti½c Evidence, and Economic Policy in an Aging World,” The Journal of the Economics of Ageing 1–2 (2013): 3–15; and from Mauricio Avendano and Lisa F. Berkman, “Labor Markets, Employment Policies, and Health,” in Social Epidemiology, 2nd ed., ed. Lisa F. Berkman, Ichiro Kawachi, and Maria Glymour (New York: Oxford University Press, 2014). 1 Social Security Administration, “Actuarial Life Table” (2010), http://www.ssa.gov/oact/ STATS/table4c6.html. 2 Vicki A. Freedman, Eileen Crimmins, Robert F. Schoeni, et al., “Resolving Inconsistencies in Trends in Old-Age Disability: Report from a Technical Working Group,” Demography 41 (3) (2004): 417–441. 3 David H. Rehkopf, Nancy E. Adler, and John W. Rowe, The Impact of Health and Education on Future Labor Force Participation of the Near Elderly in the United States, The MacArthur Founda tion Research Network on an Aging Society (forthcoming). 4 David M. Cutler, Ellen R. Meara, and Seth Richards-Shubik, “Healthy Life Expectancy: Esti- mates and Implications for Retirement Age Policy,” nber Working Paper 2011, 10–11. 5 Rehkopf, Adler, and Rowe, The Impact of Health and Education on Future Labor Force Participation of the Near Elderly in the United States. 6 Axel Boersch-Supan, Klaus Härtl, and Alexander Ludwig, “Aging in Europe: Reforms, Interna - tional Diversi½cation and Behavioral Reactions,” American Economic Review 104 (5) (2014): 1–7; James J. Heckman, Rodrigo Pinto, and Peter Savelyev, “Understanding the Mechanisms through Which an Influential Early Childhood Program Boosted Adult Outcomes,” American Economic Review 103 (2013): 2052–2086; and David N. Weil, “Accounting for the Effect of Health on Economic Growth,” Quarterly Journal of Economics 122 (3) (2007): 1265–1306.

52 Dædalus, the Journal ofthe American Academy of Arts & Sciences 7 Carlos Riumallo-Herl, Sanjay Basu, David Stuckler, Emilie Courtin, and Mauricio Avendano, Lisa F. “Job Loss, Wealth and Depression during the Great Recession in the usa and Europe,” In ter - Berkman, national Journal of Epidemiology 43 (5) (2014): 1508–1517; and Philipp Hessel and Mauricio Aven- Axel Boersch- dano, “Are Economic Recessions at the Time of Leaving School Associated with Worse Phys ical Supan & Mauricio Functioning in Later Life?” Annals of Epidemiology 23 (11) (2013): 708–715. Avendano 8 Hugo Westerlund, Mika Kivimäki, Archana Singh-Manoux, et al., “Self-Rated Health Before and After Retirement in France (gazel): A Cohort Study,” The Lancet 374 (9705) (2009): 1889–1896. 9 G. Mein, P. Martikainen, H. Hemingway, S. Stansfeld, and M. Marmot, “Is Retirement Good or Bad for Mental and Physical Health Functioning? Whitehall II Longitudinal Study of Civil Servants,” Journal of Epidemiology and Community Health 57 (1) (2003): 46–49. 10 Markus Jokela, Jane E. Ferrie, David Gimeno, et al., “From Midlife to Early Old Age: Health Trajectories Associated with Retirement,” Epidemiology 21 (3) (2010): 284–290. 11 Kerwin Ko½ Charles, “Is Retirement Depressing?: Labor Force Inactivity and Psychological Well-Being in Later Life,” in Accounting for Worker Well-Being, ed. Solomon W. Polachek (Am - sterdam; San Diego, Calif.; Oxford: Elsevier, 2004), 269–299. 12 Norma B. Coe and Gema Zamarro, “Retirement Effects on Health in Europe,” Journal of Health Economics 30 (1) (2011): 77–86. 13 Susan Rohwedder and Robert J. Willis, “Mental Retirement,” Journal of Economic Perspectives 24 (1) (2010): 119–138. 14 Beverly A. Roberts, Rebecca Fuhrer, Michael Marmot, and Marcus Richards, “Does Retirement Influence Cognitive Performance? The Whitehall II Study,” Journal of Epidemiology and Commu- nity Health 65 (11) (2011): 958–963. 15 Stéphane Adam, Eric Bonsang, Sophie Germain, and Sergio Perelman, “Retirement and Cogni - tive Reserve: A Stochastic Frontier Approach Applied to Survey Data,” crepp Working Papers 2007/04 (2007). 16 Rohwedder and Willis, “Mental Retirement.” 17 Norma B. Coe, Hans-Martin von Gaudecker, Maarten Lindeboom, and Jürgen Maurer, “The Effect of Retirement on Cognitive Functioning,” Journal of Health Economics 21 (8) (2012): 913–927. 18 Eric Bonsang, Stéphane Adam, and Sergio Perelman, “Does Retirement Affect Cognitive Func - tioning?” roa Research Memorandum 2010/1 (Maastricht, The Netherlands: Maastricht Uni - versity, 2010); Adreas Kuhn, Jean-Philippe Wuellrich, and Josef Zweimüller, “Fatal Attraction? Access to Early Retirement and Mortality,” Institute for Empirical Research Working Paper 499, Institute for Empirical Research in Economics (University of Zurich, 2010); Rohwedder and Willis, “Mental Retire ment”; and Fabrizio Mazzonna and Franco Peracchi, “Aging, Cog - nitive Abilities, and Retirement,” European Economic Review 56 (4) (2012): 691–710. 19 Esteban Calvo, Natalia Sarkisian, and Christopher R. Tamborini, “Causal Effects of Retirement Timing on Subjective Physical and Emotional Health,” The Journals of Gerontology: Psycho- logical Sciences & Social Sciences 68 (1) (2013): 73–84. 20 Stefanie Behncke, “Does Retirement Trigger Ill Health?” Health Economics 21 (3) (2012): 282–300. 21 Axel Boersch-Supan, “Myths, Scienti½c Evidence and Economic Policy in an Aging World,” The Journal of the Economics of Ageing 1–2 (2013): 3–15. 22 Jonathan Gruber and David A. Wise, eds., Social Security Programs and Retirement around the World: The Relationship to Youth Employment (Chicago: University of Chicago Press, 2010); and Axel Boersch-Supan and Reinhold Schnabel, “Early Retirement and Employment of the Young in Germany” in Gruber and Wise, eds., Social Security Programs and Retirement around the World, 147–166.

144 (2) Spring 2015 53 Labor-Force 23 Boersch-Supan and Schnabel, “Early Retirement and Employment of the Young in Germany,” Participation, in Gruber and Wise, eds., Social Security Programs and Retirement around the World, 147–166. Policies & 24 Practices in Ibid. an Aging 25 Gruber and Wise, eds., Social Security Programs and Retirement around the World. America 26 Mauricio Avendano and Ichiro Kawachi, “Why do Americans Have Shorter Life Expectancy and Worse Health than Do People in Other High-Income Countries?” Annual Review of Public Health 35 (2014): 307–325.

54 Dædalus, the Journal ofthe American Academy of Arts & Sciences Productivity & Engagement in an Aging America: The Role of Volunteerism

Dawn C. Carr, Linda P. Fried & John W. Rowe

Abstract: Volunteering in late life is associated with health bene½ts such as reduced risk of hypertension, improved self-related health and well-being, delayed physical disability, enhanced cognition, and lower mortality. Although the mechanisms of these correlations are not clear, increases in physical activity, cognitive engagement, and social interactions likely play contributing roles. Volunteers are typically thought to represent a select group, often possessing higher levels of education and income, good health, and strong social networks. However, group evidence indicates that there are many members of groups of lower socio - economic status (SES), including elderly adults, who serve their communities on a regular basis and in high-priority programs. We propose that the impact of volunteering in an aging population be recognized and invested into, and that effective programs harness social capital of older adults to address critical societal needs and also improve the well-being of older adults. While members of low-SES groups are less likely to volunteer, they exhibit disproportionately great bene½ts. The Experience Corps represents a model of an effective volunteerism program, in which elders work with young schoolchildren. Existing federal initiatives, including the Foster Grandparent Program and Senior Companion Program–which target low-income elders–have had low participation with long waiting lists. Given the proven bene½ts and relatively low proportion of older persons who volunteer, enhancement of elder volunteerism presents a signi½cant opportunity for health promotion and deserves consideration as a national public health priority.

Recent and expected future increases in life ex - pectancy and the increasing proportion of our population that will be elderly has stimulated substantial research into the factors that promote well-being DAWN C. CARR is a Social Science and health in late life. Early research on aging was Research Associate at Stanford Cen - concerned primarily with understanding the aver- ter on Longevity at Stanford Uni- age or usual physiologic and psychological changes versity. associated with aging, particularly in the context of LINDA P. FRIED is Dean of the inevitable loss and decline as part of senescence. The Mailman School of Public Health ½rst White House Conference on Aging in 1961, how- at Columbia University. ever, reoriented gerontological research to provide JOHN W. ROWE, a Fellow of the information that facilitates good societal and indi- American Academy since 2005, is vidual choices associated with positive aging out- Professor at the Columbia Universi- comes.1 More than twenty-½ve years later, John Rowe ty Mailman School of Public Health. and gerontologist Robert Kahn–working as part of (*See endnotes for complete contributor the MacArthur Foundation Research Network on an biographies.) Aging Society–sought to advance discussions about

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00330 55 Produc - successful aging to enhance our under- paid volunteering–is considered part of tivity & standing of the mechanisms involved.2 the general category of volunteering. While Engagement in an Aging They proposed that successful aging should there are de½nitional differences among America: re flect the distinction between two non - many of the major available data sources, The Role of Volun - path ologic forms of aging–usual and suc - volunteer engagement among older people teerism cessful–and called for research investigat- seems to be on the rise over the last several ing the factors underlying the heterogene- decades, with somewhere between one in ity among older people. Following a decade four and one in three older people in the of systematic studies in this area, the Mac- United States volunteering today.8 Among Arthur Network laid out three critical fac- older volunteers, approximately half ded- tors to successful aging: 1) avoidance of icate two or more hours per week on av- disease and disability; 2) maintenance of erage, with the rest involved only sporad- high cognitive and physical function; and ically.9 3) engagement with life.3 As discussed in the introduction to this A growing body of research and inter- volume, engagement through either paid ventions related to volunteering has bol- work or volunteering is an important com- stered our understanding of the range of ponent of assuring that the United States ways in which it is associated with positive will be a productive and equitable society health outcomes in later life.10 Although as it ages. A substantial body of research the health bene½ts associated with volun- indicates that remaining an active mem- teering are robust, we are only beginning ber of society through meaningful and pro- to understand the mechanisms of the pos - ductive social roles yields many bene½ts to itive bene½ts attributed to volunteer en - the elderly.4 For those concerned about gagement. the impact of aging at both the individual Three primary mechanisms have been and population levels, volunteerism is an hypothesized to produce these bene½ts: attractive area of study: it leverages human increased physical engagement, cognitive capital to create social capital, offers sub- engagement, and social interaction.11 stantial health bene½ts that facilitate a First, with respect to physical bene½ts, vol - suc cessful aging lifestyle, and facilitates so - unteering has been shown to be associated cietal cohesion as a powerful tool for con- with reduced risk of onset of diseases (in- necting generations through a shared sense cluding decreased risk of hypertension), of purpose.5 decreased mortality risk, improved self- Distinct from forms of societal engage- rated health, and delayed decline in phys- ment such as caregiving, providing infor- ical functioning.12 These physical health mal help to friends or family, or paid work, bene½ts purportedly stem, at least in part, volunteering typically refers to what is from the extent to which volunteering in- commonly understood to include working volves increased levels of physical engage- for an organization for no (or very modest) ment, though most volunteer activities in- pay in a capacity that would otherwise clude only mild or moderate levels of ac- involve ½scal remuneration.6 It also is pre - tivity. Second, with respect to cognitive sumed to be an activity in which the indi- bene½ts, volunteering has been shown to vidual involved is uncoerced and driven be related to enhanced cognitive func- primarily by a concern for his or her com- tion.13 The cognitive bene½ts attributed to munity.7 Working in exchange for very volunteering are proposed to relate to the limited compensation, often to cover trans - level of cognitive engagement required to portation or meal expenses–so-called perform the tasks associated with volun-

56 Dædalus, the Journal ofthe American Academy of Arts & Sciences teering, which include executive planning are also more likely to work at least part Dawn C. and use of memory.14 Third, with respect time), being married (particularly for Carr, Linda P. to social interaction, in addition to being those whose spouse also volunteers), and Fried & linked to decreased depressive symp- being in good health all increase the like- John W. toms,15 social engagement in volunteering lihood of volunteering: these characteris- Rowe is associated with enhanced overall well- tics provide individuals with greater capac - being, with increased bene½ts with more ity to contribute than their lower resourced time spent volunteering.16 Volunteer ac - peers, and their social connections facili- tivities typically involve social interactions, tate greater access to opportunities (for ex - with both the people whom the volunteers ample, they are more likely to be asked to are helping and those whom they are vol- volunteer).22 Research has shown that so- unteering alongside. Some researchers cial networks that value volunteer engage - have presumed that the value of increased ment produce in their members a stronger social interaction to mental health is based sense of obligation to volunteer.23 Offer- in part on the feeling of “mattering” to ing a stipend and volunteer opportunities others.17 However, the bene½ts of main- of high value to the community and espe- taining meaningful relationships with oth- cially to children are important strategies ers has been shown to have far-reaching to attract signi½cant numbers of volun- effects on longevity, and could point to teers from all racial and ethnic back- physiological factors such as decreases in grounds. Mi nority groups also spend con- overall stress.18 siderable time volunteering in church or In addition to our lack of detailed un- other community groups that are often derstanding of the mechanisms underlying “un der the ra dar” of scholars or agencies the bene½ts of volunteering, we have scant who evaluate volunteerism activities. In information on the “dose response” of the addition, sociologists Yunqing Li and Ken - bene½ts, including both the intensity and neth Ferraro found that individuals who duration of the engagement. While a few struggle with depression are more likely to studies suggest that two hours per week of seek out vol unteering opportunities, and volunteering produces the greatest ben- that they experience a decrease in symp- e½ts, with additional engagement produc- toms with formal engagement in volun- ing no additional bene½t and potentially teering.24 Under standing the impact of leading to detrimental effects on health,19 such selection ef fects on the dynamics of others suggest that engagement beyond volunteering is required to better under- two hours per week on average does, in stand the causal pathways between volun- fact, produce more signi½cant health ben - teering and well-being, particularly if we e½ts.20 Clearly the dose of volunteering are interested in maximizing the public that yields the greatest individual health health impact of volunteer engagement. bene½ts has yet to be determined, and new Volunteering is not static: people fre- insights into dose responsivity will be key quently move into or out of this form of en- to the design of volunteering initiatives. gagement. People may volunteer for many years and then stop when they experience A key potential limitation of a major ex- a health event that prevents continued en- pansion of senior volunteering relates to gagement, or they may switch from one selection effects: the signi½cant differ- organization to another that more readily ences between volunteers and nonvolun- accommodates their abilities.25 In addi- teers.21 High educational attainment, suf- tion, some individuals who have never vol- ½cient income (in part because volunteers unteered may reach later life and choose

144 (2) Spring 2015 57 Produc - to begin volunteer engagement at the re- what degree they are involved, but also tivity & quest of a friend. Although the factors how much they bene½t from it. For in - Engagement in an Aging driving these behaviors cannot be fully ex - stance, despite engaging in volunteering at America: plained using the large observational data much lower rates, individuals with lower The Role of Volun - sets for which much of the research ½nd- levels of resources have been shown to ex - teerism ings associated with volunteering are perience disproportionately higher ben - based, at a broader level, some important e½ts from volunteering.34 patterns have emerged. First, once someone becomes a volun- The abundant evidence demonstrating teer, he or she is more likely to remain one; that those who volunteer are better re- thus, it is easier to keep a volunteer from sourced and better poised to volunteer quitting than it is to get a nonvolunteer to than those who do not has raised concerns start volunteering, especially when the about volunteering being a privilege.35 If socio political context values volunteering volunteering offers evidence of individual and offers positive reinforcement.26 Sec- “success” in aging, the alternative (pre- ond, those with past volunteer experiences sumably unsuccessful or “usual” aging) are likely to return to volunteering if they may be depicted as a reflection of an indi- do stop, particularly if they have a history vidual’s poor choices.36 In other words, of volunteering at higher intensity levels.27 since volunteering would seem to be a life- Third, those who have never volunteered style choice, the onus for obtaining the ben - are less likely to start; and if they have fair e½ts of volunteering is then placed on in - or poor health, are disabled, have limita - dividuals, who may or may not have the tions related to executive function, or have means to participate. In addition, with vol - less than a high school education, they are unteering producing such potent health highly unlikely to start volunteering.28 bene½ts, the extent to which certain groups Fourth, the dynamics associated with vol - of individuals lack access to volunteer roles unteering are influenced by the dynamics in later life suggests that unequal ability of engagement in other productive activ- to participate in volunteering is a major ities. In general, by simply engaging in any public health and health-disparities con- other productive activities (including care- cern. With these considerations in mind, a giving or work) individuals are more likely new agenda for volunteerism research has to start a new volunteer role.29 Decreas- been to identify ways to minimize volun- ing time spent engaged in paid work is also teer disparities and, by extension, mini- associated with increased likelihood of mize health disparities among older adults. starting a volunteer role30 and compared As noted above, individuals with lower to those who fully retire, those who choose levels of resources have been shown to to retire into a part-time job are more experience disproportionately greater ben- likely to start or continue volunteering.31 e½ts from volunteering.37 Interest in in - Finally, despite the increase in volunteer- creasing participation in volunteering ing that accompanies work-hour reduc- among underrepresented groups has led tions, it is much more likely that a non- to several interventions designed to en - volunteer will start volunteering in later hance participation among older people. life if they begin volunteering prior to re - The ½rst step in building the interventions tirement,32 if they marry/are married to was to gain a clearer understanding of the a volunteer, or if they are asked to volun- key barriers to volunteering. These barriers teer.33 Importantly, selection effects may include issues related to disability, cost to influence not only who volunteers and to the individual, access, opportunity/incen-

58 Dædalus, the Journal ofthe American Academy of Arts & Sciences tive, and social network and environmen- tary compensations can incentivize volun - Dawn C. tal factors. Institutional level responses teer participation. Nonmonetary incen- Carr, Linda P. have effectively enhanced participation of tives include training or skill development Fried & older nonvolunteers and retained existing opportunities, but more common strate- John W. volunteers using ½ve primary strategies: gies include goods or services such as food; Rowe 1) designing for high impact of service; 2) medical services such as free physical ex - role flexibility; 3) recognition; 4) accom- aminations; gift certi½cates; or prizes.42 modation and training; and 5) compen- These strategies have been shown to be sation.38 Offering flexible volunteer op - helpful for enhancing participation some- portunities has been shown to increase what, but are minimally associated with participation, and it may be an especially sustained volunteer engagement.43 ef fective tool for encouraging underrep- Compensation was proposed to increase resented groups who may have fewer re- sustained engagement speci½cally of un - sources and who are more likely to face derrepresented groups, but researchers greater informal and formal care responsi - discovered that all older adults–regardless bilities or time-consuming and potentially of socioeconomic status–½nd the stipend unpredictable health problems. Recogniz- important. This is for two reasons: 1) most ing older adults’ contributions are also elderly are on a ½xed income, and the small im portant to increasing participation stipend covers out of pocket costs of vol- and maintaining volunteers. Older people unteering; and 2) a stipend lends credi- want to ensure that the ways their time is bility to the program, since it demonstrates being used matters, and tend to respond that organizers believe its impact is worth more strongly to positive feedback that the ½nancial investment. Programs like relates to the bene½t of younger people;39 Peace Corps and AmeriCorps have utilized this positivity enhances the effect of re - monetary stipends to increase participa- warding altruistic commitments.40 Older tion for decades, and the recently imple- people who have fewer skills and abilities mented Edward M. Kennedy Serve Amer - in later life (such as those with lower levels ica Act of 2009 called for an increase in of educational attainment) often feel less stipend volunteer roles for people of all con½dent about being able to volunteer, ages. However, ½nancial compensation for despite having a desire to do so. In addition volunteer engagement during later life was to having fewer skills, those with fewer ½rst introduced by initiatives designed to resources are also disproportionately more increase community service engagement likely to have health problems that may among low-income older adults. In 1965, limit mobility and, thus, the ability to vol - the ½rst program associated with today’s unteer. Organizations that offer training or “Senior Corps” programs was introduced: skill development necessary to successful the Foster Grandparent Program. This pro - volunteering and who are able to accom- gram was designed to provide both a modate mobility problems, including by meaningful community and intergenera- providing transportation, enjoy an increase tional engagement role for older adults, in sustained participation.41 and an income supplement for low-income The fourth and perhaps most contro- elders.44 In 1974, another federally spon- versial approach used to increase the par- sored Senior Corps program was intro- ticipation of individuals in underrepre- duced: the Senior Companion Program. sented groups in volunteering during later This program continues to offer ½nancial life is compensation, or so-called paid vol - stipends for income-eligible older people. unteering. Both monetary and nonmone- The Senior Companion Program supports

144 (2) Spring 2015 59 Produc - older adult volunteers who visit the homes the use of stipends as an intervention to tivity & of frail elders, providing them with social increase engagement of underrepresented Engagement in an Aging support and thereby enhancing the re - older adults in volunteering and to im- America: sources they have available to age in place prove the health of individuals and com- The Role of Volun - successfully. Both Senior Corps programs munities has been shown to be successful teerism were designed to encourage high engage- in a more recently introduced program: ment volunteer roles–those requiring ½f - the Experience Corps. The Experience teen to forty hours of service per week– Corps brings older adult volunteers into and support antipoverty efforts. public elementary schools to help improve These volunteer programs have been students’ academic achievement. De - min imally modi½ed since their introduc- scribed below, the Experience Corps has tion, and there are no studies assessing the observed higher recruitment rates, longer impact of the stipend or if the programs volunteer tenure, increased hours of en- decrease disparities. Because funding lev- gagement, and increased bene½ts associ- els are so limited, participation in these ated with participation with use of sti - programs continues to be fairly low. For pends.46 example, in 2012, the Foster Grandparent Program saw only 28,500 participating out The Experience Corps, the most robustly of the eligible nineteen million adults aged studied volunteer program designed for ½fty-½ve and above living at or below 200 older Americans, has clearly demonstrated percent of the poverty line. Lack of par- that volunteer interventions can play an ticipation could be related to the stigma important role in enhancing the well-being associated with receipt of a means-based of older volunteers as well as the bene½cia- stipend, poor health, lack of access to op- ries they serve. The core of the model that portunities, or due to participation in other became the Experience Corps (a title sug- roles (such as paid work or caring for fam- gested by Lyndon B. Johnson’s Secretary ily members); but the presence of long of Health, Education, and Welfare, John waiting lists for these programs suggest Gardner) was independently designed and that limited funding is playing a major role proposed by Linda Fried and Encore.org- in their failure to grow. While the Foster founder Marc Freedman, who together Grandparent Program has elicited partic- col laborated on the ½nal design in 1994. ipation from diverse older adults who are The overall strategy was to embed an evi- income eligible, should funds become dence-based health-promotion/disease- available to further expand eligibility, en- prevention program in a senior volunteer- gagement by older adults with moderate ing initiative to create a community- and moderate-to-low income may greatly based social model of high-impact health increase the public health and social im - enhancement. The hypothesis was that pact of this program. Like other means- this approach would deliver effective pre- tested programs, some individuals who vention and health promotion into the are income-eligible may not participate be- community for all older adults–including cause of the attention brought to their ½ - those who otherwise might not access nancial circumstances. Furthermore, those health promotion programs–and would unable or uninterested in high-engage- lead to decreased rates of mobility and ment volunteering are also excluded.45 iadl (instrumental activities of daily liv- Although there are discrepancies in the ing) disability, frailty, falls, and cognitive extent to which it is an effective policy and decline. The approach would also produce program tool for all volunteer programs, delayed onset of these ailments at a pop-

60 Dædalus, the Journal ofthe American Academy of Arts & Sciences ulation level, thus contributing to a com- teen hours per week throughout the full Dawn C. pression of morbidity. The vehicle through school year. They served and were trained Carr, Linda P. which this prevention would occur would in teams for the roles they would perform, Fried & have a generative impact, organizing and and learned the unique challenges associ- John W. amplifying the social capital offered by an ated with twenty-½rst-century schools. At Rowe aging society to support improved academ- the conclusion of the pilot demonstration, ic outcomes of vulnerable children and the all ½ve sites reported that the model met teaching effectiveness and ef½cacy of teach- all of its original criteria for success: pro- ers. Ultimately, the goal was to demonstrate viding roles that were of importance to that a new social institution could be de - principals and meaningful to volun- signed to create meaningful roles and re - teers; using high intensity, ½fteen-hours- sponsibilities for older adults while also per-week service; providing comprehen- exposing the bene½ts for all generations sive training; deploying volunteers in achieved through the in creased engage- teams; deploying a critical mass of volun- ment of older adults in an aging society. teers in each school; providing a stipend; The Experience Corps model targets chil- creating a diverse volunteer force; improv- dren from kindergarten to third grade, re - ing health of volunteers and building a flecting the research that suggests that chil- vehicle for gen erative impact; establishing dren who do not succeed in school by the pathways to leadership for volunteers; and third grade are more likely to drop out. ensuring that all elements of the program Congressional support was provided for be a win-win-win for children, schools, pilot studies in ½ve cities: New York (via and older adults. the Community Service Society); Phila - After the initial success of the pilot pro- delphia; Minneapolis; Port Arthur, Tex - gram, a second successful national demon- as; and Portland, Oregon. Implemented stration was launched to target literacy at through the Corporation for National and the original ½ve sites. Linda Fried, who was Community Service (with Linda Fried and then based at Johns Hopkins, led the ini- Marc Freedman), the pilot project was tiative to expand the program to Baltimore. conducted between 1996 and 1997 to as- These demonstrations were followed by sess design elements and roles for older a highly successful pilot randomized tri - adults, to identify requirements for imple- al, which was published in 2004.47 There- mentation, and to determine feasibility after, Freedman started Civic Ventures and acceptability to older volunteers and (now Encore.org) in San Francisco to create schools. Dur ing this demonstration, it be - a movement built around service by older came evi dent that it was impossible in adults and to organize a franchise of pro- some cities to recruit for this signi½cant grams, formalized as Experience Corps™. time commitment (½fteen hours per week) The program has since grown to include without of fer ing a stipend; as a result, all twenty-three cities, many of whom pro- sites started offering a stipend at the level vide funding for the program, and in 2009, that the Fos ter Grandparent Program pro - the program became af½liated with the vided: $200 per month. This potentiated aarp. Studies of the Experience Corps older adults living on modest ½xed incomes model have shown remarkable results. K–3 being able to serve by providing money for students in the intervention schools, as bus fare and other expenses of service. compared to those in control schools, have The model was carefully designed, de - shown improved standardized reading ploying a critical mass of older adults at scores and markedly fewer referrals for be - each school, all of whom committed ½f- havioral problems. The results to date ap -

144 (2) Spring 2015 61 Produc - pear to suggest that boys bene½t from the er adults would respond to the opportu- tivity & program more than do girls. Teachers and nity to volunteer for a program designed Engagement in an Aging principals report large improvements in for high impact on the futures of children. America: school atmosphere and climate (school This was conditioned on providing a mod- The Role of Volun - safety, delinquency, classroom order, learn - est monthly stipend (about $200–250 tax - teerism ing environment) with a critical mass of able dollars per month) that covered the high time-commitment Experience Corps costs of volunteering (bus fare, lunches, volunteers in the school. and so on) for older adults who had limited The bene½ts to older adult volunteers resources. This stipend served as incen- have been particularly robust.48 They re- tive to volunteer because it signi½ed that ported experiencing higher levels of social society expects the volunteer service to integration and sense of generative achieve - make a difference. It also contributed to ment than controls. Additionally, the num- full participation: with a little money on ber of hours of service was proportional to the line, volunteers seemed to be motivat- bene½ts. Overall, there were modest ben- ed to get up each morning and participate e½ts to lifestyle, intellectual, and physical when they may otherwise have decided not activity at twelve months. Perhaps most to. Perhaps more important, because of the important, those with low levels of each stipend, minority older adults became the type of activity at baseline show meaning- dominant volunteer group. ful and signi½cant increases. For physical To date, the Experience Corps demon- activity, increases were approximately 800 strates that older people will volunteer to Kcal burned per week, an amount consis- make a difference for the next generation, tent with a modest exercise program. Ex - and that a societal institution that trans- perience Corps also showed the ½rst evi- forms human capital into social capital for dence that a community-based activity en- generativity, in a model designed for high gagement program directly impacts mark - impact, can harness this energy. People ers of brain health known to buffer the par ticipate to ensure their legacy as well brain from the clinical expression of neuro- as to give back. Retention is high because pathologies, such as Alzheimer’s disease volunteers receive evidence that they are and vascular dementia. Findings indicated making a difference. The Experience Corps that length as well as dose of exposure mat- is both a volunteer and public health pro- ters: for men, the bene½ts emerge during gram, delivering a high and sustained dose the second year of service. Older women of prevention to diverse older adults: ½f- with baseline low/normal levels of cogni- teen hours per week of increased physical, tive function experienced improved exec- cognitive, and social activity and social en- utive function and corresponding brain ac- gagement/integration with meaning and tivation on fmri (functional mri) within purpose. The return on investment has one year.49 been demonstrated to be high, and could At the outset of Experience Corps, the increase dramatically when the long-term prevailing “wisdom” related to volunteer- impact on children’s and older adults’ out- ing was based primarily on upper socio - comes are assessed. economic status (ses) white women. Ex- pe rience Corps showed that levels of Unlike other health interventions that fa- informal community, civic, and church- cilitate similar health bene½ts in late life, based service in the African-American such as exercise, volunteer programs have community is signi½cant. Fried correctly the additional potential to provide a means theorized that minority and lower ses old - to address important social problems and

62 Dædalus, the Journal ofthe American Academy of Arts & Sciences thus strengthen civil society.50 With the lating reserve of knowledge, skills, and ex - Dawn C. combination of addressing important so - periences of older adults to target speci½c Carr, Linda P. cial problems and the bene½ts offered to social problems and simultaneously facili - Fried & those who engage, volunteering has gained tate compressed morbidity of the older John W. attention among policy-makers in the population. Rowe United States and in other developed and With these considerations in mind, and developing nations. In addition to positive in view of the accumulated scienti½c evi- contributions to the health of older adults dence of its bene½ts, the time has come to –potentially facilitating delayed onset of identify late-life engagement through vol- morbidity and mortality–volunteering unteering as a major public health issue, has also been shown to leverage human with special emphasis on engagement of capital in a way that facilitates workforce in dividuals across the full socioeconomic opportunities. Particularly for those who spectrum, including those with fewer re - have less ½nancial resources in retirement, sources who have the most to gain. This enhancing opportunities to maintain in - effort will require federal and local sup- come in later life is critical. Volunteering port, as well as additional research to iden- increases available social resources by fa - tify both the speci½c “dose” at which vol- cilitating social network connections and unteering yields the maximum bene½t and opportunities to obtain skills that are valu- the best strategies to recruit individuals able in the paid workforce. As we prepare with diverse social characteristics. If suc- for U.S. society to age successfully, volun- cessful, a comprehensive national effort to teer engagement programs should play an enhance volunteerism in late life can be an important role. The Experience Corps important component of our successful dem onstrates that a volunteer interven- transition to a productive and equitable tion can successfully leverage the accumu- aging society.

endnotes * Contributor Biographies: DAWN C. CARR is a Social Science Research Associate at Stanford Center on Longevity at Stanford University. She edited Gerontology in the Era of the Third Age: Implications and Next Steps (with Kathrin Komp, 2011); has contributed articles to such journals as Activities, Adaptation and Aging, The Gerontologist, and Journal of Health and Human Services Administration; and serves on the editorial board of The Gerontologist. LINDA P. FRIED is Dean of the Mailman School of Public Health, Senior Vice President of the Columbia University Medical Center, the DeLamar Professor of Public Health, and Professor of Epidemiology and Medicine at Columbia University. She is the designer and cofounder of Experience Corps, and also founded the Johns Hopkins Center on Aging and Health. She serves on the editorial boards of The Journals of Gerontology: Biological Sciences & Medical Sciences and Aging. JOHN W. ROWE, a Fellow of the American Academy since 2005, is Professor at the Columbia University Mailman School of Public Health and Chair of the MacArthur Foundation Research Network on an Aging Society. He is the author of Successful Aging (with Robert L. Kahn, 1998) and was the Chair of the Institute of Medicine of the National Academies project the Future Health Care Workforce for Older Americans, which authored the report Retooling for an Aging America: Building the Health Care Workforce (2008).

144 (2) Spring 2015 63 Produc - 1 Robert J. Havighurst, “Successful Aging,” The Gerontologist 1 (1) (1961): 8–13, doi:10.1093/ tivity & geront/1.1.8. Engagement in an Aging 2 John W. Rowe and Robert L. Kahn, “Human Aging: Usual and Successful,” Science 237 (4811) America: (1987): 143–149, doi:10.1126/science.3299702. The Role 3 of Volun - Robert L. Kahn, “Guest Editorial on ‘Successful Aging and Well-Being: Self-Rated Compared teerism with Rowe and Kahn,’” The Gerontologist 42 (6) (2002): 725–726, doi:10.1093/geront/42.6.725; and John W. Rowe, “The New Gerontology,” Science 278 (5337) (1997): 367–375, doi:10.1126/ science.278.5337.367. 4 Caroline E. Jenkinson, Andy P. Dickens, Kerry Jones, Jo Thompson-Coon, Rod S. Taylor, Morwenna Rogers, Clare L. Bambra, Iain Lang, and Suzanne H. Richards, “Is Volunteering a Public Health Intervention? A Systematic Review and Meta-Analysis of the Health and Survival of Volunteers,” BMC Public Health 13 (2013): 773–783, doi:10.1186/1471-2458-13-773; Ming-Ching Luoh and A. Regula Herzog, “Individual Consequences of Volunteer and Paid Work in Old Age: Health and Mortality,” Journal of Health and Social Behavior 43 (4) (2002): 490–509; Yunkyung Jung, Tara L. Gruenewald, Teresa E. Seeman, and Catherine A. Sarkisian, “Productive Activities and Development of Frailty in Older Adults,” The Journals of Geron- tology: Psychological Sciences & Social Sciences 65 (2) (2010): 256–261, doi:10.1093/geronb/ gbp105; Nicole D. Anderson, Thecla Damianakis, Edeltraut Kröger, Laura M. Wagner, Deirdre R. Dawson, Malcolm A. Binns, Syrelle Bernstein, Eilon Caspi, Suzanne L. Cook, and the bravo Team, “The Bene½ts Associated with Volunteering among Seniors: A Critical Review and Recommendations for Future Research,” Psychological Bulletin 40 (6) (2014): 1503–1533, doi:10.1037/a0037610; Hui-Xin Wang, Anita Karp, Bengt Winblad, and Laura Fratiglioni, “Late-Life Engagement in Social and Leisure Activities is Associated with Decreased Risk of Dementia: A Longitudinal Study from the Kungsholmen Project,” American Journal of Epi- demiology 155 (12) (2002): 1081–1087, doi:10.1093/aje/155.12.1081; Thomas A. Glass, Carlos F. Mendes De Leon, Shari S. Bassuk, and Lisa F. Berkman, “Social Engagement and Depressive Symptoms in Late Life Longitudinal Findings,” Journal of Aging and Health 18 (4) (2006): 604–628, doi:10.1177/0898264306291017; Jesse E. Hinterlong, Nancy Morrow-Howell, and Philip A. Rozario, “Productive Engagement and Late Life Physical and Mental Health: Findings from a Nationally Representative Study,” Research on Aging 29 (4) (2007): 348–370; Terry Y. Lum and Elizabeth Lightfoot, “The Effects of Volunteering on the Physical and Mental Health of Older People,” Research on Aging 27 (1) (2005): 31–55, doi:10.1177/0164027504271349; Nancy Morrow-Howell, “Volunteering in Later Life: Research Frontiers,” The Journals of Gerontology: Psychological Sciences & Social Sciences 65 (4) (2010): 461–469, doi:10.1093/geronb/gpq024; and Jan E. Mutchler, Jeffrey A. Burr, and Francis G. Caro, “From Paid Worker to Volunteer: Leaving the Paid Workforce and Volunteering in Later Life,” Social Forces 81 (4) (2003): 1267–1293, doi:10.1353/sof.2003.0067. 5 Generations United and the Generations Initiative, Out of Many, One: Uniting the Changing Faces of America (Washington, D.C.: Generations United, 2013), http://www.gu.org/RESOURCES/ Publications/OutofManyOne.aspx; Jean E. Kincade, Donna J. Rabiner, Shulamit L. Bernard, Alison Woomert, Thomas R. Konrad, Gordon H. DeFriese, and Marcia G. Ory, “Older Adults as a Community Resource: Results from the National Survey of Self-Care and Aging,” The Geron - tologist 36 (4) (1996): 474–482, doi:10.1093/geront/36.4.474; and Greg O’Neill, Sarah F. Wilson, and Nancy Morrow-Howell, “The Civic Enterprise: Advancing Civic Engagement Opportu- nities in Later Life,” in Civic Engagement in an Older America, ed. Greg O’Neill and Sarah F. Wilson (Washington, D.C.: Gerontological Society of America, 2010). 6 Morrow-Howell, “Volunteering in Later Life.” 7 Ram A. Cnaan, Femida Handy, and Margaret Wadsworth, “De½ning Who is a Volunteer: Conceptual and Empirical Considerations,” Nonpro½t and Voluntary Sectory Quarterly 25 (3) (1996): 364–383. 8 Barbara A. Butrica, Richard W. Johnson, and Sheila R. Zedlewski, “Volunteer Dynamics of Older Americans,” The Journals of Gerontology: Psychological Sciences & Social Sciences 64 (5) (2009): 644–655, doi:10.1093/geronb/gbn042.

64 Dædalus, the Journal ofthe American Academy of Arts & Sciences 9 Morrow-Howell, “Volunteering in Later Life.” Dawn C. 10 Carr, Jenkinson et al., “Is Volunteering a Public Health Intervention?”; Anderson et al., “The Bene½ts Linda P. Associated with Volunteering among Seniors”; Ben Heaven, Laura J.E. Brown, Martin White, Fried & Linda Errington, John C. Mathers, and Suzanne Moffatt, “Supporting Well-Being in Retirement John W. through Meaningful Social Roles: Systematic Review of Intervention Studies,” The Milbank Rowe Quarterly 91 (2) (2013): 222–287, doi: 10.1111/milq.12013; and Linda P. Fried, Michelle C. Carlson, Marc Freedman, Kevin D. Frick, Thomas A. Glass, Joel Hill, Sylvia McGill, George W. Rebok, Teresa Seeman, James Tielsch, Barbara A. Wasik, and Scott Zeger, “A Social Model for Health Promotion for an Aging Population: Initial Evidence on the Experience Corps Model,” Journal of Urban Health: Bulletin of the New York Academy of Medicine 81 (1) (2004): 64–78, doi:10.1093/jurban/jth094. 11 Anderson et al., “The Bene½ts Associated with Volunteering among Seniors”; and Fried et al., “A Social Model for Health Promotion for an Aging Population.” 12 Luoh and Herzog, “Individual Consequences of Volunteer and Paid Work in Old Age”; Jung et al., “Productive Activities and Development of Frailty in Older Adults”; Lum and Light- foot, “The Effects of Volunteering on the Physical and Mental Health of Older People”; Morrow-Howell, “Volunteering in Later Life”; and Mutchler et al., “From Paid Worker to Volunteer.” 13 Anderson et al., “The Bene½ts Associated with Volunteering among Seniors”; and Wang et al., “Late-Life Engagement in Social and Leisure Activities is Associated with Decreased Risk of Dementia.” 14 Michelle C. Carlson, Kirk I. Erickson, Arthur F. Kramer, Michelle W. Voss, Natalie Bolea, Michelle Mielke, Sylvia McGill, George W. Rebok, Teresa Seeman, and Linda P. Fried, “Evi- dence for Neurocognitive Plasticity in At-Risk Older Adults: The Experience Corps Program,” The Journals of Gerontology: Biological Sciences & Medical Sciences 64 (12) (2009): 1275–1281, doi:10.1093/gerona/glp117. 15 Glass et al., “Social Engagement and Depressive Symptoms in Late Life Longitudinal Find- ings”; Hinterlong, Morrow-Howell, and Rozario, “Productive Engagement and Late Life Phys- ical and Mental Health”; and Lum and Lightfoot, “The Effects of Volunteering on the Physical and Mental Health of Older People.” 16 Heaven et al., “Supporting Well-Being in Retirement through Meaningful Social Roles”; and Fried et al., “A Social Model for Health Promotion for an Aging Population.” 17 Jane Allyn Piliavin and Erica Siegl, “Health Bene½ts of Volunteering in the Wisconsin Lon- gitudinal Study,” Journal of Health and Social Behavior 48 (4) (2007): 450–464, doi:10.1177/ 002214650704800408. 18 Seoyoun Kim and Kenneth F. Ferraro, “Do Productive Activities Reduce Inflammation in Later Life? Multiple Roles, Frequency of Activities, and C-Reactive Protein,” The Gerontologist 54 (5) 2014: 830–839. 19 Luoh and Herzog, “Individual Consequences of Volunteer and Paid Work in Old Age”; Lum and Lightfoot, “The Effects of Volunteering on the Physical and Mental Health of Older People”; and Marieke Van Willigen, “Differential Bene½ts of Volunteering Across the Life Course,” The Journals of Gerontology: Psychological Sciences & Social Sciences 55 (5) (2000): S308–S318, doi:10.1093/geronb/55.5.S308. 20 Nancy Morrow-Howell, Jim Hinterlong, Philip A. Rozario, and Fengyan Tang, “Effects of Vol - un teering on the Well-Being of Older Adults,” The Journals of Gerontology: Psychological Sciences & Social Sciences 58 (3) (2003): S137–S145, doi:10.1093/geronb/58.3.S137; and Rodlescia S. Sneed and Sheldon Cohen, “A Prospective Study of Volunteerism and Hypertension Risk in Older Adults,” Psychology and Aging 28 (2) (2013): 578–586, doi:10.1037/a0032718. 21 Peggy A. Thoits and Lyndi N. Hewitt, “Volunteer Work and Well-Being,” Journal of Health and Social Behavior 42 (2) (2001): 115–131, doi:10.2307/3090173.

144 (2) Spring 2015 65 Produc - 22 Morrow-Howell, “Volunteering in Later Life”; Butrica, Johnson, and Zedlewski, “Volunteer tivity & Dynamics of Older Americans”; Dawn C. Carr and Jon Hendricks, “Relevance of Social Cap - Engagement ital and Lifestyle for the Third Age,” in Gerontology in the Era of the Third Age: Implications and in an Aging , ed. Dawn C. Carr and Kathrin S. Komp (New York: Springer, 2011): 207–224; Tay America: Next Steps The Role K. McNamara and Ernest Gonzales, “Volunteer Transitions among Older Adults: The Role of of Volun - Human, Social, and Cultural Capital in Later Life,” The Journals of Gerontology: Psychological Sci- teerism ences & Social Sciences 66 (4) (2011): 490–501; and Carroll L. Estes, Jane L. Mahakian, and Tracy A. Weitz, “A Political Economy Critique of ‘Productive Aging,’” in Social Policy and Aging: A Crit- ical Perspective, ed. Carroll L. Estes (Thousand Oaks, Calif.: Sage Publications, 2001), 187–200. 23 Butrica, Johnson, and Zedlewski, “Volunteer Dynamics of Older Americans.” 24 Yunqing Li and Kenneth F. Ferraro, “Volunteering in Middle and Later Life: Is Health a Bene½t, Barrier or Both?” Social Forces 85 (1) (2006): 497–519. 25 Butrica, Johnson, and Zedlewski, “Volunteer Dynamics of Older Americans.” 26 Mutchler, Burr, and Caro, “From Paid Worker to Volunteer.” 27 Francis G. Caro and Scott A. Bass, “Receptivity to Volunteering in the Immediate Postretire- ment Period,” Journal of Applied Gerontology 16 (4) (1997): 427–441, doi:10.1177/07334648970160 0403; and Fengyan Tang, Nancy Morrow-Howell, and Eunhee Choi, “Why Do Older Adult Volunteers Stop Volunteering?” Ageing and Society 30 (5) (2010): 859–878. 28 Butrica, Johnson, and Zedlewski, “Volunteer Dynamics of Older Americans.” 29 Ibid. 30 Mutchler, Burr, and Caro, “From Paid Worker to Volunteer.” 31 Dawn C. Carr and Ben Lennox Kail, “The Influence of Unpaid Work on the Transition Out of Full-Time Paid Work,” The Gerontologist 53 (1) (2013): 92–101. 32 Sheila R. Zedlewski, “Will Retiring Boomers Form a New Army of Volunteers?” The Retirement Project: Perspectives on Productive Aging (7) (2007): 1–7. 33 Butrica, Johnson, and Zedlewski, “Volunteer Dynamics of Older Americans.” 34 Morrow-Howell, “Volunteering in Later Life.” 35 Estes, Mahakian, and Weitz, “A Political Economy Critique of ‘Productive Aging.’” 36 Kahn, “Guest Editorial on ‘Successful Aging and Well-Being’”; and William J. Strawbridge, Margaret I. Wallhagen, and Richard D. Cohen, “Successful Aging and Well-Being: Self-Rated Compared with Rowe and Kahn,” The Gerontologist 42 (6) (2002): 727–733. 37 Morrow-Howell, “Volunteering in Later Life.” 38 Fengyan Tang, Nancy Morrow-Howell, and Song-Iee Hong, “Inclusion of Diverse Older Pop - ulations in Volunteering: The Importance of Institutional Facilitation,” Nonpro½t and Voluntary Sector Quarterly 38 (5) (2008): 810–827, doi:10.1177/0899764008320195. 39 Laura L. Carstensen, “Social and Emotional Patterns in Adulthood: Support for Socioemo- tional Selectivity Theory,” Psychology and Aging 7 (3) (1992): 331–338. 40Morrow-Howell et al., “Effects of Volunteering on the Well-Being of Older Adults.” 41 Andrea Galiette Skoglund, “Do Not Forget about Your Volunteers: A Qualitative Analysis of Factors Influencing Volunteer Turnover,” Health and Social Work 31 (3) (2006): 217–220. 42 Ram A. Cnaan and Toni Cascio, “Performance and Commitment: Issues in Management of Vol- unteers in Human Service Organizations,” Journal of Social Service Research 24 (3/4) (1998): 1–37. 43 Tang, Morrow-Howell, and Choi, “Why Do Older Adult Volunteers Stop Volunteering?” 44 Rosalyn Saltz, “Aging Persons as Child-Care Workers in a Foster Grandparent Program: Psycho - social Effects and Work Performance,” The International Journal of Aging and Human Develop- ment 2 (4) (1971): 314–340.

66 Dædalus, the Journal ofthe American Academy of Arts & Sciences 45 Fengyan Tang, Nancy Morrow-Howell, and Song-Iee Hong, “Inclusion of Diverse Older Pop - Dawn C. ulations in Volunteering: The Importance of Institutional Facilitation,” Nonpro½t and Voluntary Carr, Sector Quarterly 38 (5) (2008): 810–827, doi:10.1177/0899764008320195. Linda P. Fried & 46 Tang, Morrow-Howell, and Choi, “Why Do Older Adult Volunteers Stop Volunteering?”; and John W. Amanda Moore McBride, Ernest Gonzales, Nancy Morrow-Howell, and Stacey McCrary, Rowe “Stipends in Volunteer Civic Service: Inclusion, Retention, and Volunteer Bene½ts,” Public Administration Review 71 (6) (November/December 2011): 850–858. 47 Fried et al., “A Social Model for Health Promotion for an Aging Population.” 48Song-Iee Hong and Nancy Morrow-Howell, “Health Outcomes of Experience Corps: A High- Commitment Volunteer Program,” Social Science and Medicine 71 (2) (2010): 414–420. 49Carlson et al., “Evidence for Neurocognitive Plasticity in At-Risk Older Adults.” 50 Caroline E. Jenkinson, Andy P. Dickens, Kerry Jones, Jo Thompson-Coon, Rod S. Taylor, Morwenna Rogers, Clare L. Bambra, Iain Lang, and Suzanne H. Richards, “Is Volunteering a Public Health Intervention? A Systematic Review and Meta-Analysis of the Health and Survival of Volunteers,” BMC Public Health 13 (2013): 773–783, doi:10.1186/1471-2458-13-773.

144 (2) Spring 2015 67 Resetting Social Security

S. Jay Olshansky, Dana P. Goldman & John W. Rowe

Abstract: Social Security retirement bene½ts were ½rst introduced in 1935 as a ½nancial safety net for a large and rapidly growing older American population. The program was intended to be economically self- sustaining, but population aging and rising life expectancies threaten the program’s solvency. The 1983 Social Security Amendments mandated that the full retirement age increase to 67 by the year 2027. In this essay, we present evidence demonstrating that the rate of improvement in life extension at older ages accelerated after 1983. If the 1935 ratio of working years to retired years is maintained, early and full retirement ages of 66.5 and 69.4, respectively, were justi½ed in 2009. Additional delays in the age of eligibility beyond those currently in effect would place signi½cant ½nancial burdens on individuals with lower life ex pectancies, the poor and near-poor, and the very old, and–absent additional reform–would exacerbate existing unequal access to entitlements within the system.

In the future when there are a great many persons over 65, most of the able-bodied individuals will and should continue working to age 70 or 75 if their services seem needed. –Robert J. Myers, Chief Actuary (1947–1970) and Deputy Commissioner (1981–1982) of the Social Security Admin- istration (ssa) and leader of the National Commission on Social Security Reform (1982–1983)1 S. JAY OLSHANSKY is Professor of Epidemiology at the School of Public Health, Division of Epide - Has the time arrived to reset the age of eligibility miology and Biostatistics at the for Social Security retirement bene½ts? When Pres- Uni versity of Illinois at Chicago. ident Roosevelt signed the Social Security Act (ssa) in 1935 in the wake of the Great Depression, unem- DANA P. GOLDMAN is Professor of Public Policy, Pharmacy, and ployment was 34 percent, savings accounts were dec - Economics at the University of imated, and almost 50 percent of the older popula- Southern California. tion was dependent on family and friends for ½nancial support. There was reason to believe large seg- JOHN W. ROWE, a Fellow of the ments of the population–particularly the elderly– American Academy since 2005, is 2 Professor at the Columbia Universi- were facing destitution. ty Mailman School of Public Health To address this concern, the Committee on Eco- and Chair of the MacArthur Foun- nomic Security was established by executive order dation Research Network on an in 1934. What we know today as Social Security began Aging So ciety. simply as a federally administered social insurance (*See endnotes for complete contributor retirement program for older people, nominally ½- biographies.) nanced through payroll taxes and paid for by work-

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00331 68 ers and their employers. As the program Social Security tax rates.6 Coverage was S. Jay was originally structured by the Social Se - also extended to dependents of retired Olshansky, Dana P. curity Act of 1935, people would earn ben - work ers or workers who died prematurely. Goldman e½ts as they continued to work. If death In 1948, bene½ts to dependents, survivors, & John W. occurred before age sixty-½ve, or before and those with severe and long-lasting dis- Rowe they received what they paid into the sys- ability were increased or extended and cov- tem even after retirement, their estate erage was expanded considerably.7 In 1950, would receive the difference plus interest a revised schedule of gradual increases in in the form of a one-time lump-sum pay- tax rates for employers and employees was ment. At the program’s inception, no ben- implemented to increase the likelihood e½ts were provided to spouses or children. that Social Security would remain self- Although Social Security was originally supporting; coverage was also extended designed to protect a limited number of to several additional major categories of American workers against loss of earnings, workers such as farmers and government President Roosevelt indicated from the workers.8 Legislation in 1954 and 1956 ex - start that the program was expected to tended coverage to 90 percent of all work- grow and evolve with changing economic ers, and coverage became nearly universal and demographic conditions.3 The ½rst in the early 1960s.9 The eligibility age for study published by the Of½ce of the Actu- Social Security was reduced from age 65 to ary at the Social Security Board claimed age 62 for women in 1956 and for men in that “when it is realized that too large a 1961, and automatic cost-of-living adjust- proportion of the population would prob - ments were authorized in 1972.10 Finally, ably be left idle with a retirement age of 65, in direct response to gains in life expectan- the general feeling will undoubtedly be cy and improvements in health (increases that a constant retirement age should be in active and disability-free–or what we banished, or that it should be left as a bal- prefer to call “healthy”–life expectancy) ancing item.”4 A subsequent publication since the program began, amendments ap - by Robert J. Myers, Chief Actuary and Dep - proved in 1983 authorized gradual increas- uty Commissioner of the ssa–from which es in the age of full eligibility for workers we quote in our epigraph–made a more born after 1937, with provisions fully ef- forceful statement about raising the Social fective for all workers born after 1959.11 Security Retirement age.5 These amendments gradually increased Social Security has evolved extensively the age of eligibility for full Social Security since its inception. While the program is bene½ts from 65 to 67 and lowered the ben - best known for providing ½nancial assis- e½ts for those who choose to begin receiv- tance to retirees, amendments to the pro- ing them early (between 62 and the full re - gram also added life insurance, payments tirement age). There have been no lon - for spouses and dependents, and disability gevity- or health-related adjustments to bene½ts for those who are unable to work the retirement age since 1983. It is also im - but are not yet eligible by age for regular portant to emphasize that, today, approx- bene½ts. The ½rst signi½cant change to the imately 72 percent of new bene½ciaries program was introduced in 1939, when draw bene½ts before the full retirement age Congress passed amendments to change and 46 percent draw bene½ts at the earliest the ½nancing of the program so that work- possible age of 62.12 Despite the program’s ers paid into Social Security incrementally evolution, therefore, the question remains as they worked, allowing for immediate whether eligibility changes have kept pace payments of bene½ts without increasing with the substantial gains in life expec -

144 (2) Spring 2015 69 Resetting tancy and healthy life expectancy that have indexed directly to rising life expectancy Social occurred since the program’s inception since the program’s inception, maintaining Security and, indeed, since the last retirement-age a constant proportion of adult life spent adjustment in 1983. working to life spent in retirement?14 The current debate about raising the age of ½rst eligibility above 62 and the age of Improvements in health care and increas- full-bene½t eligibility above 67 has been es in well-being at older ages have accel- driven by a combination of factors, includ - erated in the United States since Social Se- ing: ½nancial stress placed on the solvency c urity began in 1935 with a set retirement of the Social Security trust fund by a much age of 65. At that time, the average expec- larger number of bene½ciaries than antic- ted remaining years of life for someone ipated (in turn caused by an unexpected reaching age 65–notated as e(65)–for men baby boom and larger-than-anticipated in- and women combined in the United States creases in life expectancy); a substantial was 12.6 years, and the probability of sur- proportion of bene½ciaries who elect early viving to age 65 (averaged for men and bene½ts; political reluctance to increase women) conditional on having survived to payroll taxes; and a growing number of age 25 (referred to as “conditional sur- very long-living older people who depend vival”) was 62.4 percent (Table 1). By 1983, fully or nearly so on Social Security (called e(65) for the total population had risen to “longevity risk”). Today, two-thirds of 16.6 years (meaning that each year during bene½ciaries rely on Social Security for this time frame, 30 days were added to the more than half of their total income, and life of a person reaching 65 years of age), 25 percent rely on it for over 90 percent of while conditional survival to age 65 rose their total income.13 The shift toward re - to 79.4 percent. Between 1983 and 2009, tirees relying fully on the program for ½- life expectancy past 65 rose an additional nancial support was neither anticipated 2.3 years to 18.9, which means that the nor intended at the program’s inception. annual increase in life expectancy accel- Taken together, these considerations erated to 31.8 extra days added to the life lead to the four central questions we ad- of a 65-year-old per year; conditional sur- dress in this essay: vival to age 65 also increased to 84.8 per- 1) How well did the two-year increase in cent between 1983 and 2009. eligibility age for full retirement bene½ts Since many bene½ciaries now retire at from the 1983 amendments correspond to the earliest possible retirement age of 62, the proportional rise in life expectancy at it is worth noting that e(62) increased by 4.3 age 65 from 1935 to 1983? years between 1935 and 1983, and by 2.5 2) From a demographic perspective, does years between 1983 and 2009 (see Table 1). the rise in life expectancy at older ages This means 32 additional days of life were observed since 1983 warrant a further ad - added each year to those reaching age 62 justment to the age of eligibility for early from 1935 to 1983, and 35.2 additional days and full Social Security bene½ts? of life were added each year for those 3) How would subgroups of the U.S. pop- reaching age 62 from 1983 to 2009. ulation with different survival prospects Conditional survival to the full Social be differentially influenced by further in- Security retirement age of 65 varies consid- creases in the age of early and full retire- erably by sex and level of completed educa- ment ages? tion; trends in conditional survival be - 4) And what would the early and full re- tween 1990 and 2008 reveal large differenc - tirement ages be today if they had been es among population subgroups (Table 2).

70 Dædalus, the Journal ofthe American Academy of Arts & Sciences Table 1 S. Jay U.S. Life Expectancy at Age 62 [e(62)], Life Expectancy at Age 65 [e(65)], and Olshansky, Conditional Survival from Age 25 to Age 65 [S ], by Sex; 1935, 1983, 2009 Dana P. (25–65) Goldman & John W. e(62) e(65) S(25–65) Rowe M F T M F T M F T 1935 13.6 15.1 14.4 11.9 13.2 12.6 59.5 67.3 63.3 1983 16.3 20.9 18.7 14.3 18.6 16.6 74.8 85.7 80.2 2009 19.7 22.6 21.2 17.5 19.9 18.9 81.9 88.7 85.3

M = Male; F = Female; T = Total (average). Source: U.S. Social Security Administration, Historical Life Tables developed by the Of½ce of the Chief Actuary for use in estimates and analysis in The 2013 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds (2014); see http://www.ssa.gov/oact/TR/2013/tr2013.pdf.

Table 2 Percent of Total U.S. Population Surviving to Age 65 Conditional on Having Survived to Age 25, by Level of Completed Education; 1990, 2008

Years of Education at Age 25 <12 12 13–15 16+ 1990 75.7 78.3 88.1 86.0 2008 74.4 78.7 89.2 92.1

Source: Calculations done by the MacArthur Foundation Research Network on an Aging Society.

In 1990, only 75.7 percent of 25-year-old not live long enough to draw retirement men and women with less than a high bene½ts from Social Security at current eli- school education were expected to reach gibility ages. In contrast, only 5.9 percent age 65. In contrast, about 87 percent of the of the most educated group will die before most highly educated 25-year-olds in that the early retirement age. year were expected to survive to age 65. In The observed full retirement age of 67 2008, the least educated experienced a mandated in 1983 will not be implemented slight reduction in survival to age 65 (down until 2027. If the full retirement age had to 74.4 percent) while the most highly ed - been indexed exclusively to e(65) (that is, if ucated experienced a signi½cant additional the full retirement age was raised in pro- improvement (to 92.1 percent). Condition- portion to the increase in life expectancy al survival increased from 1990 through at age 65 using a 10-year moving average), a 2008 as a function of level of completed full retirement age of 67.7 would have been education; the biggest jump in survival oc- justi½ed in 1983 (see Figure 1). If the full re - curred among those who have any college tirement age was indexed again in 2009 to education. Thus, 25.6 percent of the least life expectancy at age 65, a full re tirement educated subgroup of the population will age of 69.4 would have been jus ti½ed. And

144 (2) Spring 2015 71 Resetting Figure 1 Social Projected Full Retirement Age (Based on a 10-Year Moving Average) Indexed to a Constant Ratio of Security Retired to Working Years Present in 1935, Statutory Full Retirement Age (United States, 1935–2009), and Estimated Average Retirement Age for Males and Females Combined (United States, 1950–2005)

72 Projected full retirement age indexed to a constant ratio of retired to working years present in 1935 (10-year moving average) 70

66 Age at Retirement 64

Statutory full 62 retirement age

Estimated average retirement age (males and females combined) (15) 60 1935 1983 2009 Year

Source: M. Gendell, “Retirement Age Declines Again in 1990s,” Monthly Labor Review 124 (2001): 12–21.

if in 2009 the full retirement age had been dexed early retirement age has also fallen indexed to either observed life expectancy further behind the statutory early retire- at birth or conditional sur vival to age 65, a ment age every year since it began in 1956, full retirement age above 70 would be jus - and this too accelerated after 1983. ti½ed. The indexed full retirement ages The full retirement age for Social Secu- have not only been falling further behind rity is currently applied equally to almost the statutory retirement age every year everyone in the population, regardless of since 1935 (a difference whose growth ac- individual attributes that differentially celerated after 1983), but the ac tual (esti- in fluence survival prospects or monetary mated) average age at full retirement in the contributions and expenses. How would United States has been steadily declining the retirement age vary if it were linked to well below both the indexed and stat utory the observed longevity attributes of pop- full retirement ages ever since 1965. ulation subgroups? Here, we use the ex - Along the same lines, indexing the early ample of education attainment (the latest retirement age to the rise in life expectancy data available are for 2008). after age 62 (using a ten-year moving aver- If the full retirement age had been in - age) observed since 1956 (when the amend- dexed in 2008 to e(65) for population sub- ment lowering age of eligibility to age 62 groups demarcated by level of completed was ½rst enacted for women) would have education (relative to the standard that ex- justi½ed an early retirement age of 64.8 in isted for the total population in 1935), the 1983 and 66.5 in 2009 (see Figure 2). The in- full retirement age in 2008 would have

72 Dædalus, the Journal ofthe American Academy of Arts & Sciences Figure 2 S. Jay f Projected Early Retirement Age (Based on a 10-Year Moving Average) Indexed to a Constant Ratio of Olshansky, Retired to Working Years Present in 1935 (United States, 1935–2009), and Statutory Early Retirement Dana P. Goldman Age (1956–2009) & John W. Rowe 72

70 Projected early retirement age indexed to a constant ratio of retired to working years present in 1935 (10-year moving average) 68 Statutory early retirement age

66 Age at Retirement 64

60 1935 1983 2009 Year

Source: Figure prepared based on Social Security Administration life tables.

been 68.8 for the least educated and 71.8 for ysis above, conducted in 2004, would have the most educated. The early retirement yielded an early retirement age of 66.3. age would have been 66.1 for the least educated and 69.6 for the most educated. When determining the full Social Secu - Another consideration that is often rity retirement age in 1935, the Council on raised in the retirement-age debate is the Economic Security (ces) based the ½g- possibility of “correcting” for enhance- ure of 65 on retirement ages commonly ments in life expectancy through stabili- used in private pension systems of the zation of the ratio of years working to years time. About half of state retirement retired.15 In 1935, assuming a full retire- schemes used 65 as the retirement age, ment age of 65, the population aged 20 while the other half used 70. The ces set- and older spent 78 percent of its remaining tled on age 65 in part due to actuarial life life working. If we were to hold this ratio tables suggesting that the lower age would of working to retired years constant and yield a manageable self-sustaining system index the full retirement age to rising life based on modest levels of payroll taxa- expectancy at age 65, the full retirement tion.16 The question now is not whether age would have been 69.1 years in 2009 the most appropriate retirement age was (based on a ten-year moving average in life chosen in 1935, but what the appropriate expectancy after 65; see Figure 1). Assum- ages would be from a demographic per- ing an early retirement age of 62, the over- spective if the ages for early and full re - 20 population in 1935 spent 74 percent of tirement were adjusted to how long we ac - their remaining life working; and the anal- tually live.

144 (2) Spring 2015 73 Resetting When the ssa was passed into law in because the change in retirement age did Social 1935, government actuaries were acutely not begin to take effect until 2006 and the Security aware of the forthcoming demographic full increase to 67 will not be realized until shift of population aging, and they antic- 2027, the implementation of the change ipated not only an increase in the size of was too slow. the bene½ciary population, but also a rise 2) Does the rise in life expectancy at older ages in longevity, improved survival rates to age observed since 1983 warrant a further adjust- 65, and improvements in health.17 One ac- ment to the age of eligibility for early and full tuarial report even mentions how aging Social Security bene½ts, and if so, what would science could, in the future, delay senility they be? Evidence indicates that the rate and prolong life far beyond what the ssa of improvement in survival past age 65 then considered a normal lifespan. 18 How - accelerated between 1983 and 2009, and ever, their initial predictions were still too does warrant an actuarially justi½ed full conservative. Population projections ini- retirement age (based on a 10-year moving tially generated by ssa actuaries antici- average) of 69.4 and an early retirement pated that the proportion of the total U.S. age of 66.5 in 2009. population aged 65 and older would nev- 3) How would subgroups of the U.S. popu- er exceed 15 percent.19 As of 2014, it is 14 lation with diverse survival prospects be differ- percent, but it is expected to rise up to 20 entially influenced by further increases in early percent by mid-century.20 It was also and full retirement ages? Social Security re- projected that 20 million bene½ciaries (at tirement bene½ts are made available to the very most) would ever draw bene½ts everyone that contributed to the pro- from Social Security after the year 2000;21 gram over a suf½cient period of employ- and that the size of the 65-and-older pop- ment. The amount paid out to bene½cia- ulation in the United States would peak at ries is linked to their level of contribu- about 31 million in 2025 (under the most tion, but there is a cap on the maximum optimistic mortality scenario).22 Today, and a floor on the minimum amount that there are more than 43 million people in they can receive. Although the age at the United States aged 65 and older.23 Fur- which Social Security retirement outlays ther more, Social Security was originally may begin is not influenced by gender, en visioned as a supplementary form of re - occupation, level of completed education, tirement income designed as a safeguard or health status, all of these attributes (and against destitution; it was never anticipat- others) create considerable var iation in ed that so many people would become fully both the prospect of living long enough to ½nancially dependent on the program. reach retirement ages, and how many years With these demographic shifts in mind, of life people have remaining after retire- we will summarize some of our ½ndings ment. and answer the questions raised in the in - By way of example, of those who began troduction. working at age 25 in the United States in 1) How well did the two-year increase in eli- 2009, 88.8 percent of women and 81.3 gibility age for full retirement bene½ts from the percent of men are projected to survive to 1983 amendments correspond to the propor- age 65. Of those, 32.6 percent of women tional rise in life expectancy from 1935 to 1983? and 19.5 percent of men will reach their Evidence presented here indicates that ninetieth birthday.24 At opposite ends of the 1983 amendments raising the full re - the longevity spectrum, about 32 percent tirement age to 67 effectively anticipated of men with less than a high school edu- the increase in longevity since 1935, but cation will die before age 65, whereas only

74 Dædalus, the Journal ofthe American Academy of Arts & Sciences 6 percent of women with a college degree healthy life expectancy has changed over S. Jay will die before age 65 (as of 2008).25 time.28 Nonetheless, this much is clear: Olshansky, Dana P. When the Social Security full-retirement many more people survive to age 65 today Goldman age of 65 was chosen in 1935, it was based than did in 1935; the observed increase in & John W. on the observed longevity experience of the bene½t-collecting population is far Rowe the white population in the United States– greater than originally anticipated; those the subgroup with the highest life expec- reaching ages 65 and older are now living tancy.26 Although this was done to create much longer than was ever thought pos- the most favorable payout to retirees, it sible; and there has been a notable increase also created an immediate disparity for the in healthy life expectancy by people reach - nonwhite working population as well as ing older ages today relative to any other for populations of all races with less than generation in American history.29 twelve years of education and/or in pov - There is no formula currently available erty: fewer members of these population to utilize health-related data to guide fur- subgroups were expected to live long ther adjustments to the full and early re - enough to retire, and those that did reach tirement ages; and even if there were, it retirement had shorter post-retirement life would be dif½cult to create policy respon- expectancies. Although the size of this less sive to the fact that people move in and out educated subgroup has been declining in of states of health and disability as they the United States since 1935, the “retire- age. Currently, if healthy life expectancy ment-bene½t disparity” has been rising be- were used in any way to guide adjustments cause the longevity of the most educated to the retirement ages, they could only sup - has been increasing at a faster pace than port increases above those already defen- that of the least educated.27 sible from rising longevity alone (see Fig- Any increases in the early and full retire- ures 1 and 2). However, any formula link- ment ages (including those already being ing improved health to increases in the implemented under the 1983 amendments) early and full retirement ages should also exacerbate economic disadvantage among take into account the possibility that the those less likely to reach retirement ages– health of the population in general, and a disadvantage that continues after retire - future older cohorts in particular, could ment, since those with shorter survival grow worse in the coming decades.30 prospects will draw bene½ts for less time. 4) What would the early and full retirement One strategy to compensate for this effect ages be today had they been indexed directly to would be to enhance the disability pro - rising life expectancy since Social Security’s in - gram so that individuals who are unable to ception, holding constant the 1935 proportion work and are caught in the lengthy pro - of adult life spent working to life in retirement? cess of applying for disability as they reach We have determined that an early retire- age 62 (when they would have received re - ment age of 66.5 and a full retirement age tirement bene½ts) could more quickly and of 69.4 are justi½ed based on today’s life easily receive disability payments. expectancy. The idea of indexing the fu- The criteria that served as the basis for ture retirement age to observed life expec- raising the full retirement age in 1983 were tancy and the ratio of retired to working improved longevity and health. However, years has appeared in European and U.S. healthy life expectancy did not become a academic and government publications.31 routinely reported vital statistic in the The Commission on Fiscal Responsibility United States until the 1970s, so it is dif - and Reform (Simpson-Bowles) recently ½cult to make a de½nitive claim about how rec ommended that the full retirement

144 (2) Spring 2015 75 Resetting age be gradually raised to 68 by 2050 and of the early actuarial advisors for the pro- Social to 69 by 2075; and that the early retire- gram, who foresaw that the retirement age Security ment age be raised to 63 and 64, accord- would evolve in response to changing dem - ingly. This would be accomplished by in - ographic, health, and economic condi- creasing the retirement age by 1 month tions; the modi½cations we suggest are every 2 years, beginning in 2027 when the also consistent with the observed retire- current set of increases expire. ment ages of half of the states that had re- Estimates of the linkage between sur- tirement laws in effect as far back as 1935. vival and retirement ages provided here In short, in order for retirement ages to indicate that the proposal by the Simpson- “catch up” to observed levels of longevity Bowles Commission underestimated the in the United States (improvements in magnitude of the increase in the retire- health notwithstanding), an adjustment to ment ages required to keep up with ob - the 1983 amendments would be required. served longevity. The Commission’s pro- Delays in both the full and early retirement posed full retirement ages of 68 for 2050 age of ½ve months per year would be re - and 69 for 2075 should have been instated quired, beginning in 2015 and continuing in 1987 and 2005, respectively, according for thirty years. to our analysis. The Commission’s pro- However, as unambiguous as the actu- posed early retirement ages of 63 for 2050 arial and health-related justi½cations are, and 64 for 2075 should have gone into ef- further research on the circumstances sur- fect in 1955 and 1976, respectively. The rise rounding aging and retirement is neces- in life expectancy already observed for the sary, including social, economic, and po - U.S. population from 1935 to 2009 indi- l itical analyses. We acknowledge that any cates that the early and full retirement additional delay in the age of eligibility ages today should be well above the retire- beyond those already in effect from the ment ages proposed by the Commission 1983 amendments would place a signi½- to occur some sixty-one years from now. cant new ½nancial burden on subgroups of the population and exacerbate dramat- Our conclusions about resetting the ically the unequal access to entitlements early and full retirement ages for Social that is already present in the system.32 Security are based exclusively on actual There are numerous other ways in which changes in longevity observed since 1935 the Social Security trust fund can be made and do not reflect an economic or political solvent for future generations, including analysis. Using only the two criteria for de - increased taxes and revenues and reduc- termining age at full retirement that were tions in bene½ts. Equally important, the considered by the Council on Economic progressivity of our safety net cannot be Security in 1935–actuarial justi½cation and viewed through the lens of just one pro- ½nancial soundness of the program–we gram. Reforming retirement programs ½nd that there is now justi½cation to reset might also necessitate changes in disability the early and full Social Security retirement assistance and programs such as the Af - ages several years higher than they current- fordable Care Act and Medicare. Whatever ly are. Using the two criteria of improved the policy solution, this analysis highlights longevity and health since the 1983 amend- the impact that changing longevity can ments as the basis for adjusting retirement have on the progressivity and durability of ages, there is further justi½cation for such government assistance. a change. Such modi½cations in eligibility age would be consistent with the vision

76 Dædalus, the Journal ofthe American Academy of Arts & Sciences endnotes S. Jay S JAY OLSHANSKY Olshansky, * Contributor Biographies: . is Professor of Epidemiology at the School of Pub- Dana P. lic Health, Division of Epidemiology and Biostatistics at the University of Illinois at Chicago. Goldman He has published articles in such journals as The New England Journal of Medicine, JAMA: The Jour - & John W. nal of the American Medical Association, Science, The Scientist, Scienti½c American, and Health Affairs. Rowe DANA P. GOLDMAN is the Leonard D. Schaeffer Chair in Health Policy at the University of Southern California. He is also the Director of the Schaeffer Center for Health Policy and Economics. He serves as a health policy adviser to the Congressional Budget Of½ce, and his research has appeared in the New England Journal of Medicine, JAMA, Demography, the Journal of the American Statistical Association, and Health Affairs. JOHN W. ROWE, a Fellow of the American Academy since 2005, is Professor at the Columbia University Mailman School of Public Health and Chair of the MacArthur Foundation Research Network on an Aging Society. He is the author of Successful Aging (with Robert L. Kahn, 1998) and was the Chair of the Institute of Medicine of the National Academies project the Future Health Care Workforce for Older Americans, which authored the report Retooling for an Aging America: Building the Health Care Workforce (2008). 1 Robert J. Myers, “A Comparison of Dependent and Productive Groups in Various Populations,” Actuarial Study No. 2, Of½ce of the Actuary, Social Security Board, 1938, 18. 2 Martha A. McSteen, “Fifty Years of Social Security,” U.S. Social Security Administration, 2014, http://www.ssa.gov/history/50mm2.html. 3 Ibid. 4 W. R. Williamson and Robert J. Myers, “Cost Estimate for Various Modi½cations of the Old- Age Insurance Bene½ts under Title II Suggested to the Social Security Board,” Actuarial Study No. 1, Of½ce of the Actuary, Social Security Board, 1937, 8. 5 Myers, “A Comparison of Dependent and Productive Groups in Various Populations,” 18. 6 McSteen, “Fifty Years of Social Security.” 7 Ibid. 8 Ibid. 9 Ibid. 10 Ibid. 11 U.S. Social Security Administration, “Social Security Fact Sheet: Increase in Retirement Age,” 2014, http://www.ssa.gov/pressof½ce/IncRetAge.html. 12 Eric Kingson and Monique Morrissey, “Can Workers Offset Social Security Cuts by Work- ing Longer?” Economic Policy Institute Brie½ng Paper No. 343, 2012, http://www.epi.org/ publication/bp343-social-security-retirement-age/. 13 U.S. Social Security Administration, “Basic Facts,” http://www.ssa.gov/news/press/basicfact .html 14 Life expectancy and survival estimates for males and females in the United States in 1935, 1983, and 2009 used in this essay are based on life tables developed by the Of½ce of the Chief Actuary, http://www.ssa.gov/oact/TR/2013/tr2013.pdf. Life expectancy and survival estimates for the U.S. population in 2008 by age, sex, and level of completed education were drawn from complete life tables derived from the research of the MacArthur Foundation Re search Network on an Aging Society. See S. Jay Olshansky, Toni Antonucci, Lisa Berkman, Robert H. Binstock, Axel Boersch-Supan, John T. Cacioppo, Bruce A. Carnes, Laura L. Carstensen, Linda P. Fried, Dana P. Goldman, James Jackson, Martin Kohli, John Rother, Yuhui Zheng, and John Rowe, “Differences in Life Expectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up,” Health Affairs 31 (8) (2012): 1803–1813. These tables were generated from the Centers for Disease Control and Prevention, National Center for Health

144 (2) Spring 2015 77 Resetting Statistics, Mortality Multiple Cause Files –2008, http://www.cdc.gov/nchs/data_access/ Social vitalstatsonline.htm. Estimates of the early and full retirement ages indexed to rising life Security expectancy since the program’s inception, holding constant the 1935 proportion of adult life spent working to years in retirement, were generated with the following formulas: R(e,x) = 20 + ((42+e(62,x)) * 0.7438) Where R(e,x) = forecast of early retirement age in year x; 20 = conditional survival to age 20; 42 = years working, de½ned as early retirement age (62) minus age at which work begins (age 20); e(62,x) = life expectancy at age 62 in year x; and 0.7438 = the proportion of life after age 20 in 1935 spent working, assuming early retirement occurred at age 62. And R(f,x) = 20 + ((45+e(65,x)) * 0.7802) Where R(f,x) = forecast of early retirement age in year x; 20 = conditional survival to age 20; 45 = years working, de½ned as full retirement age (65) minus age at which work begins (age 20); e(65,x) = life expectancy at age 65 in year x; and 0.7802 = the proportion of life after age 20 in 1935 spent working, assuming a full retirement occurred at age 65. 15 Axel Boersch-Supan, “Rational Pension Reform,” The Geneva Papers 32 (2012): 430–446, doi: 10.1057/palgrave.gpp.2510149; and National Commission on Fiscal Responsibility and Reform, “The Moment of Truth,” 2010, http://www.½scalcommission.gov/sites/½scalcommission .gov/½les/documents/TheMomentofTruth12_1_2010.pdf. 16 U.S. Social Security Administration, “Age 65 Retirement, The German Precedent,” http://www .socialsecurity.gov/history/age65.html (accessed February 12, 2015). 17 Myers, “A Comparison of Dependent and Productive Groups in Various Populations.” 18 L. O. Shudde and G. E. Immerwahr, “Old-Age and Survivors Insurance: Analysis of Long- Range Cost Factors,” U.S. Social Security Administration, Actuarial Study No. 21, 1946. 19 Myers, “A Comparison of Dependent and Productive Groups in Various Populations.” 20 U.S. Department of Health and Human Services, Administration for Community Living, “Administration on Aging (aoa): Projected Future Growth of the Older Population,” 2014, http://www.aoa.gov/Aging_Statistics/future_growth/future_growth.aspx. 21 Robert J. Myers, “New Cost Estimates for the Old-Age and Survivors Insurance System, with the Assumption of a Static Future Wage Level,” U.S. Social Security Administration, Actu- arial Study No. 17, 1942. 22 Robert J. Myers, “Illustrative U.S. Population Projections, 1946,” U.S. Social Security Admin- istration, Actuarial Study No. 24, 1948. 23 U.S. Department of Health and Human Services, Administration for Community Living, “Ad - ministration on Aging (aoa): Projected Future Growth of the Older Population.” 24 U.S. Social Security Administration, Historical Life Tables developed by the Of½ce of the Chief Actuary for use in estimates and analysis in The 2013 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds (2014); see http://www.ssa.gov/oact/TR/2013/tr2013.pdf. 25 Olshansky et al., “Differences in Life Expectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up.” 26 Robert J. Myers, “An Analysis of Bene½ts and the Progress of the Old-Age Reserve Account under Title II of the Social Security Act,” U.S. Social Security Administration, Actuarial Study No. 8, 1938.

78 Dædalus, the Journal ofthe American Academy of Arts & Sciences 27 Olshansky et al., “Differences in Life Expectancy Due to Race and Educational Differences S. Jay are Widening, and Many May Not Catch Up.” Olshansky, 28 Dana P. Eileen M. Crimmins and Yasuhiko Saito, “Trends in Healthy Life Expectancy in the United Goldman States, 1970–1990: Gender, Racial, and Educational Differences,” Social Science and Medicine 52 & John W. (2001): 1629–1641. Rowe 29 Vicki Freedman, Brenda Spillman, Patti Andreski, Jennifer C. Cornman, Eileen Crimmins, Ellen Kramarow, James Lubitz, Linda Martin, Sharon Merkin, Robert F. Schoeni, Teresa Seeman, and Timothy Waidmann, “Trends in Late-Life Activity Limitations in the United States: An Update from Five National Surveys,” Demography 50 (2) (2013): 661–671. 30 Crimmins and Saito, “Trends in Healthy Life Expectancy in the United States, 1970–1990”; and Olshansky et al., “Differences in Life Expectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up.” 31 National Commission on Fiscal Responsibility and Reform, “The Moment of Truth”; and Boersch-Supan, “Rational Pension Reform.” 32 Anna Zajacova, Jennifer Karas Montez, and Pamela Herd, “Socioeconomic Disparities in Health Among Older Adults and the Implications for the Retirement Age Debate: A Brief Report,” The Journals of Gerontology: Psychological Sciences & Social Sciences 69 (6) (2014): 973–978.

144 (2) Spring 2015 79 Global Population Aging: Facts, Challenges, Solutions & Perspectives

David E. Bloom, David Canning & Alyssa Lubet

Abstract: The rapid aging of populations around the world presents an unprecedented set of challenges: shifting disease burden, increased expenditure on health and long-term care, labor-force shortages, dissaving, and potential problems with old-age income security. We view longer life spans, particularly longer healthy life spans, as an enormous gain for human welfare. The challenges come from the fact that our current institutional and social arrangements are unsuited for aging populations and shifting demographics; our proposed solution is therefore to change our institutions and social arrangements. The ½rst section of this essay provides a statistical overview of global population aging and its contributing factors. The second section outlines some of the major challenges associated with widespread population aging. Finally, the third section of the essay describes various responses to these challenges, both current and prospective, facing individuals, businesses, institutions, and governments.

We are in the midst of an unprecedented transi- DAVID E. BLOOM, a Fellow of the tion in global demography. The world’s population American Academy since 2005, is is aging rapidly, and older adults compose a larger the Clarence James Gamble Pro- proportion of the world’s population than ever fessor of Economics and Demog- before–a share that will only increase over the raphy in the Department of Global next century. By 2050, the percentage of the United Health and Population at the Har- vard T.H. Chan School of Public States’ population that is aged sixty years and older Health. will grow from the current ½gure of about 20 per- DAVID CANNING cent to 27 percent. The global number of centenar- is the Richard ians worldwide–those aged one hundred years and Saltonstall Professor of Population Sciences and Professor of Econom- older–is expected to more than double by 2030, 1 ics and International Health in the with projections of nearly 3.4 million by 2050. Three Department of Global Health and major factors are driving this transition: decreasing Population at the Harvard T.H. fertility, increasing longevity, and the aging of large Chan School of Public Health. population cohorts. ALYSSA LUBET is a Research As- Falling fertility rates are the main determinant of sistant in the Department of Global population aging. Low fertility rates lead to smaller Health and Population at the Har- youth cohorts, which create an imbalance in the age vard T.H. Chan School of Public structure: older age groups become larger than their Health. younger counterparts. Thanks to accessible and ef - (*See endnotes for complete contributor fective birth control, increased child survival, and biographies.) cultural changes, birth rates have dropped dramat-

© 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00332 80 ically in the past century. In 1950, the glob- percent of the world’s population was sixty David E. al total fertility rate (tfr), or the average years or older; this number increased to Bloom, David number of children per woman, was about 11 percent by 2010. Over the next several Canning 5; by 2010, that number had dropped by decades, this proportion is expected to rise & Alyssa 50 percent. By 2050, the tfr will have dramatically, reaching a projected 21.2 Lubet dropped even further to about 2.25 chil- percent by 2050. The change is even more dren per woman. In many countries, fer- dramatic for the share of the world’s pop- tility rates are now well below the long- ulation aged eighty years or older. This term replacement rate of just over two chil- proportion climbed from just 0.6 percent dren per woman. in 1950 to 1.6 percent in 2010, and is pro- Changes in fertility rate are accompanied jected to make up 4.1 percent of the global y by increased longevity, another driver of population by 2050. t population aging. Averaging for sex and While the population of virtually every r location, a child born in 1950 had a life ex - country is aging rapidly, there remains pectancy of only forty-seven years, while considerable variation at both regional and d an adult who had survived to the age of country levels, with strong correlations e sixty could expect to live another fourteen to differing income levels. mdcs trend g years. In contrast, by 2010, life expectancy toward low fertility and high longevity, and at birth had increased to seventy years, and less-developed countries (ldcs) trend to- continued life expectancy for those aged ward the opposite. At the low end of the sixty increased to twenty years. In a num- fertility range are the mdcs found in Eu- ber of populations, recent increases in lon - rope and East Asia, with Bosnia, Herze- gevity have been attributed to falling rates govina, and Singapore tied for the lowest of tobacco consumption, as well as im - tfr of 1.28 children per woman. Mean- provements in medical technologies.2 By while, Sub-Saharan Africa has a regional 2050, life expectancy at birth is expected to tfr of just over 5, while also hosting the have risen to nearly seventy-seven years, highest country-level fertility rates: So - while life expectancy at age sixty will in - malia (6.61), Mali (6.86), and Niger (7.58). crease to twenty-two-and-a-half years. As for longevity, Japan is in the lead with Meanwhile, large population cohorts, a current life expectancy at birth of eighty- such as the United States’ postwar baby three-and-a-half years, in stark contrast boom generation, are moving through to Sierra Leone, where life expectancy at middle age and older adulthood. This birth is slightly over forty-½ve years. move ment can be seen in Figure 1, which Tables 1 and 2 depict the percent of the depicts the population of more-developed elderly population in the world’s most countries (mdcs) broken down by sex and and least population-aged countries, now age group. Males are on the left side of the (2010) and projected in the future (2050). pyramid and females are on the right. The The 2050 ½gures are based on a medium shifting shape of the population pyramid fertility projection, which assumes that fer- between the years 2010 and 2050 illustrates tility in all major areas will stabilize at re- the baby boom cohort’s movement from placement level (at slightly over two chil- middle into older ages. dren per woman). This comparison reveals These global phenomena–decreasing stark differences in age pro½les between fertility, increasing longevity, and the aging countries. For example, currently 23 per- of large birth cohorts–combine to drive cent of Germany’s population is aged sixty- up the percentage of older adults as a share ½ve years and older, while the correspond- of the global population. In 1950, only 8 ing ½gure for Qatar (with its large expa-

144 (2) Spring 2015 81 Global Figure 1 Population Population Pyramids for More-Developed Countries, 2010 and 2050 Aging: Facts, Challenges, 100+ 2010 Solutions & 95-99 Perspectives 90-94 85-89 80-84 75-79 70-74 65-69 60-64 55-59 50-54 Age FEMALE 45-49 40-44 MALE 35-39 30-34 25-29 20-24 15-19 10-14 5-9 0-4 1086420246810 Population (%)

100+ 2050 95-99 90-94 85-89 80-84 75-79 70-74 65-69 60-64 55-59 50-54 Age FEMALE 45-49 40-44 MALE 35-39 30-34 25-29 20-24 15-19 10-14 5-9 0-4 1086420246810 Population (%)

Source: United Nations Department of Economic and Social Affairs Population Division, Population Estimates and Projections Section, World Population Prospects: The 2012 Revision (New York: United Nations Department of Economic and Social Affairs, 2014), http://esa.un.org/unpd/wpp/index.htm.

triate worker population) is only 1 percent. labor supply, and economic growth. The These rankings are projected to shift con- economic and social consequences of siderably in the next half century, with only great er numbers and increased shares of Japan holding over in the top ½ve most the elderly will be seen in rich and poor population-aged nations. countries alike. Nations with swiftly aging populations Rapid population aging is accompanied may ½nd themselves with a growing dis- by several distinctive challenges in health, ease burden on their hands: nearly one-

82 Dædalus, the Journal ofthe American Academy of Arts & Sciences Table 1 David E. The World’s Most and Least Population-Aged Countries, 2010: Actual Population Percentages Bloom, David Canning & Alyssa Country % of population aged 65+ % of population aged 80+ Lubet Top 5 Japan 23 6 Germany 21 5 Italy 20 6 Greece 19 5 Latvia 18 4 Bottom 5 Kuwait 2 0.2 Eritrea 2 0.2 Bahrain 2 0.3 Qatar 1 0.1 United Arab 0.3 0.1 Emirates

Table 2 The World’s Most and Least Population-Aged Countries, 2050: Projected Population Percentages (Medium Fertility Model)

Country % of population aged 65+ % of population aged 80+ Top 5 Japan 37 16 Republic of 35 14 Korea Spain 35 13 Portgugal 34 12 s Cuba 34 15 f Bottom 5 Chad 4 0.4 Somalia 3 0.4 Timor-Leste 3 0.7 Mali 3 0.1 Niger 2.5 0.3

144 (2) Spring 2015 83 Global quarter of the world’s burden of disease postulates that increasing life expectan- F Population is attributable to illness in adults aged sixty cies will result not only in deaths at later Aging: Facts, and over. In turn, the majority (nearly 70 ages, but also in fewer years of life lived in Challenges, percent) of the older-adult disease burden the presence of disease and reduced phys - Solutions & Perspectives is due to noncommunicable diseases ical and cognitive functioning, resulting (ncds) such as heart disease, cancer, in healthier, as well as extended, old age.6 chron ic respiratory disease, musculo - So far, analyses of existing data to test the skeletal conditions, and mental disorders compression of morbidity hypothesis have such as Alzheimer’s and dementia.3 Ad- not been conclusive. Some studies have ding to and signi½cantly complicating shown that certain populations–such as the concerns posed by ncds is the issue of centenarians, adults with active cognitive multimorbidity, which affects a majority lifestyles and social connections, and those of older adults with ncds.4 The increasing with healthy lifestyles–do indeed seem burden of these health problems reflects to experience fewer years of illness and the epidemiological transition that has tak - disability at the end of their lives.7 How- en place over the last century and that is ever, other recent analyses found that in still occurring in many developing nations, the United States, disease prevalence has de½ned by a fundamental shift in the pre- increased along with average lifespan, and dominant causes of morbidity and mor- that years of life lived with disease and loss tality away from infectious diseases and of mobility function have increased along malnutrition and toward ncds. with life expectancies.8 Meanwhile, a The growing ncd burden could also be number of studies using data from low- a signi½cant path through which popula- and middle-income countries show no evi- tion aging slows economic growth. The dence of compression of morbidity; con- treatment and care of people suffering versely, an expansion of morbidity is also from ncds weighs heavily on government a possibility, threatening increased bur- expenditure and household wealth and dens on governments, health systems, and also results in decreased investment. In - households.9 It is clear that more research deed, where losses are quanti½able, the into the compression of morbidity will be projected economic cost of ncds is stag- necessary and that, in any case, healthy liv - gering, particularly in low- and middle- ing must be emphasized.10 income countries: recent projections show Another economic challenge presented that India stands to lose US$4.58 trillion, by population aging is the falling labor while China stands to lose US$23.03 trillion supply. In many countries, labor-force par - due to ncds in the period between 2012 ticipation falls off drastically at older ages and 2030; during this time each country’s (see Figure 2). As a result, population aging proportion of adults aged sixty-½ve and may slow national economic growth, re- older is expected to double.5 On an indi- duce asset values, strain existing pension vidual level, ncds prevent people from and health care systems, and weigh down working as long, hard, and productively as younger generations in the process. In the they otherwise might; this is reflected in United States, labor-force participation decreased labor participation rates at older rates for both sexes peak between the ages ages (see Figure 2). of forty and forty-four: in 2010, the rate for A key factor in determining the effects this group was 82.3 percent. From there, of population aging is the “compression of rates drop gradually along with age, before morbidity,” predicted in the 1980s by pro- falling precipitously from 72 percent for fessor of medicine James Fries. His theory adults aged ½fty-½ve to ½fty-nine to 55 per-

84 Dædalus, the Journal ofthe American Academy of Arts & Sciences Figure 2 David E. Labor-Force Participation Rate by Age Group, 2010 Bloom, David Canning 100% & Alyssa 90% Lubet 80% 70% 60% 50% Developed 40% Developing 30% Least Developed

- 20% 10%

Labor Force Participation Rate Participation Force Labor 0% 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Age Group

Source: International Labour Organization, ILOSTAT Database (Geneva: International Labour Organization, 2014), http://www.ilo.org/ilostat.

cent for those aged sixty to sixty-four (un - increases. This can be avoided if people fortunately, after age sixty-½ve, labor-force save real assets for their own retirement; participation rates are generally no longer but rather than accumulating real re- available in ½ve-year bands as they are for sources, many government pension sys- younger age groups).11 Similar drops can tems promise pensions based on future tax also be seen in less-developed, but still receipts. As well as money transfers to the aging, countries. In India, the participation elderly, there are often larger transfers in rate of the sixty to sixty-four age group is the form of publicly provided access to nearly 50 percent, a dramatic drop from health care, which if not ½nanced through the 64 percent participation rate of those savings must be funded through a tax bur - aged ½fty-½ve to ½fty-nine. In contrast, the den on younger workers. more population-aged Japan sees a falloff Another challenge posed by population from 80 percent to 61 percent participation aging is the prospect of slowed economic between the ½fty-½ve to ½fty-nine and six- growth by way of diminished labor and ty to sixty-four age groups. While this de - lower savings rates. There are strong life- crease reflects a greater change in percent - cycle patterns in work and saving, and age points, higher overall proportions also older generations do not work and save as point to longer working lives for a greater much as younger adults do. One dire pre- number of individuals. diction is that population aging will slow The low labor-force participation of the or perhaps even reverse the engines of na- elderly means that their consumption is tional economic growth. Reduced labor ½nanced out of either government pen- sup ply due to population aging may result sions, family transfers, or their own sav- in economies having to pay “dividend” ings. A dif½culty with transfers such as pay- back in the form of health care, long-term as-you-go pensions or informal transfers care, and capital deaccumulation as the from children to their elderly parents is elderly seek resources to ½nance their con - that they may become unsustainable as the sumption in old age. Economies may also ratio of elderly to working age population be burdened by increased social protec-

144 (2) Spring 2015 85 Global tion expenditures, such as increasing pen - respond to population aging through be - Population sion costs. Indeed, public pension expen- havioral changes in this period. Greater Aging: gdp Facts, diture as a share of is projected to rise longevity leads to longer retirements and Challenges, steeply in the coming decades due to pop- increased incentives to save during work- Solutions & Perspectives ulation aging and to earlier changes in pen- ing years in anticipation of retirement. sion coverage and bene½t rates.12 There is some evidence that people in aging societies have already adjusted to this re - The problems outlined above may be off- ality. Figure 3 shows that, for all countries set by accompanying demographic devel- in the aggregate, savings as a percentage of opments, especially if countries are pre- gdp rises along with the share of a coun- pared to take advantage of these opportu- tries’ population aged sixty-½ve years or nities. One challenge posed by population older.15 In economic terms, savings trans- aging is an altered age dependency ratio, lates into investment, which fuels the ac - or the ratio of people of working-age (ages cumulation of physical and human capi- ½fteen to sixty-four) to young children and tal and technological progress, which are older adults in a population. However, in- the classic drivers of economic growth. creased elderly dependency will be offset Individuals and households may also rein many economies by a reduced youth spond to population aging through in - dependency ratio. The elderly dependency creased investment in human capital, such ratio in the United States, for example, has as through education and training. Even grown from seventeen adults aged sixty- while a more aged population can lead to ½ve and older per one hundred working- a smaller workforce, investment in edu- aged adults in 1980 to twenty-one elderly cation can make this workforce more ef - per one hundred working-aged in 2013. In fective. Lower fertility rates lead to fewer the same time period, the ratio of children children per family, and these children under age ½fteen to working-age adults has are typically healthier and better educat- decreased from thirty-four per one hun- ed. Healthy, well-educated children gen- dred to twenty-nine per one hundred.13 erally grow up to be more productive Between 1980 and the present, the United adults.16 A workforce with higher human States’ overall age dependency ratio has capital has the potential to lead to increased re mained relatively constant, contracting productivity, wages, and standards of liv- from 51 percent to 50 percent.14 With a ing.17 Other human-capital investments lower youth dependency burden, invest- in health will generally also lead to more ment can be redirected from social spend - productive working adults, offsetting the ing on children to investment in physical reduction in the labor force as adults age capital, research and development, and in- and fertility rates decline. frastructure–all classic drivers of econom- While reduced fertility rates have shrunk ic growth. workforces in some countries, lower fer- Individuals may also respond to popu- tility has also facilitated greater labor-force lation aging through behavioral changes, participation by women. In that sense, such as increased rates of saving, higher low er fertility is tantamount to an increase educational attainment in anticipation of in the effective labor force. This will fur- longer lives, and increased labor-force par - ther offset the negative effects of popula- ticipation from women and the elderly. tion aging on workforce numbers. Older The typical “working lifespan” between people may also choose to work beyond the the ages of ½fteen and ½fty-nine is the statutory retirement age, further mitigat- prime period for saving, and people may ing this challenge. In the United States,

86 Dædalus, the Journal ofthe American Academy of Arts & Sciences Figure 3 David E. Savings by Age in Developed and Developing Countries Bloom, David Canning 60 & Alyssa Income group: Lubet High Low & Middle 40 All

0 2007-2011

-20

-40 Average Gross Domestic Savings (% of GDP), Average

-60 0 5 10 15 20 25 Population Aged 65+ (% of Total)

Source: David E. Bloom, Somnath Chatterji, Paul Kowal, Peter Lloyd-Sherlock, Martin McKee, Bernd Rechel, Larry Rosenberg, and James P. Smith, “Macroeconomic Implications of Population Ageing and Selected Policy Responses,” The Lancet 385 (9968) (2015): 649–657.

labor-force participation rates of older in - worker wellness programs as a way of in - dividuals have been increasing for the past vesting in employee health to cut down on two decades, especially among those with health care and absenteeism costs.19 Firms higher levels of schooling.18 may also take advantage of new business In addition to individuals and house- opportunities that will accompany popu- holds, businesses can respond to popula- lation aging, such as the design and mar- tion aging by adjusting human resource keting of products and services geared to - protocols and implementing technological ward older adults. innovations to assist and incentivize older With respect to public policy, it is natural workers. As older people make up greater for people to respond to longer and health- numbers of the workforce, businesses can ier lifespans by planning on longer work- shift human resources practices to meet ing lives. But since most of the world’s so - older employees’ needs for flexible roles cial security systems create strong incen- and schedules. Other developments may tives for retirement between the ages of include reallocating more physically de- sixty and sixty-½ve, public policy has been manding tasks to younger employees, of - extremely sluggish in adapting to new fering opportunities for continuing edu- dem o graphic realities. For example, data cation of older employees, and instituting on public pension systems in twenty-three

144 (2) Spring 2015 87 Global European countries between 1965 and 2005 “the growing needs of an ageing popula- Population show that while male life expectancy in tion” by way of new policies geared to- Aging: 23 Facts, those countries increased by an average of ward the health of older adults. In devel- Challenges, seven years, the mean legal retirement oping countries, especially those in which Solutions & 20 Perspectives age did not change. Fortunately, sever- family and social structures are undergo- al countries–including France, Ireland, ing rapid transformations, there is a great Greece, Sweden, and the United Kingdom need for developing basic packages of cost- –have recently raised the normal legal re- effective health services suited to the needs tirement age or have increased incentives of older people, including a realignment to delay retirement. of primary health care programs to match Some countries have also considered ad- changing demographic and epidemiolog- justing their pension systems. For exam- ical patterns.24 There is also opportunity ple, in Norway, new cohorts of older peo- to reform health care ½nancing mecha- ple will receive a pension calculated as the nisms to ensure greater fairness and sus- accumulated entitlement divided by a life tainability while also promoting risk pool- expectancy indicator. Thus, as life expec- ing and increasing ef½ciency. This has the tancy increases, the annual pension will potential to reduce the ½scal pressures as- decrease. 21 In addition, some countries are sociated with an older population, and will moving toward fully funded systems in improve the lives of older individuals by which contributions are saved in real assets providing access to more and better ser - that generate future pension income, rath - vices. In more-developed countries, a great - er than simply transferring contributions er concern is securing coverage for the of the young to ½nance pension receipts of costs and services associated with long- the old. Countries could also encourage and term care. Governments could consider re- complement behavioral shifts by investing ducing reliance on costly institutional care in schooling that will enlarge the effective by promoting self-care, in-home caretaker labor force or by emphasizing healthy liv- training, and other services that would ening and disease prevention throughout able older people to remain in their own life.22 homes. Institutional changes are also needed Reforming health education and re - to address the new demographic realities. search is an important health systems– Government-initiated policies and educa - related solution. Currently, most medical tional programs to promote ½nancial lit- curricula focus on disease cures and spe- eracy among older adults may help them cialization. A potential source of cost sav- make better choices about the forms, ac - ings and increased quality of life is to re - cessibility, and security of their assets. Old- orient medical education toward a holistic er people must make these choices while emphasis on prevention and early detec- facing uncertainty about individual longev - tion, especially given the increased burden ity and the availability of government ben - of ncds and the prevalence of multimor- e½ts; furthermore, they must make these bidity. An additional focus on healthy liv- choices during a phase of life that is asso- ing and a general understanding of co - ciated with reduced cognitive function. morbidities, treatment interactions, and Health system reform also has great po - palliative and end-of-life care, as well as tential to mitigate the negative effects of research on the physical and cognitive population aging; in 2002, the United Na - transitions associated with aging, would tions’ Madrid International Plan of Action on meet the needs of the growing elderly pop- Ageing called on governments to recognize ulation.

88 Dædalus, the Journal ofthe American Academy of Arts & Sciences On a larger scale, international migra- tions–including Japan, Korea, and those David E. tion policies have the potential to amelio- in the European Union–run their own on - Bloom, David rate the economic effects of population ag - going sister studies, allowing for interna- Canning ing, insofar as youthful developing-coun- tional harmonization and comparisons. & Alyssa try populations can ½ll job vacancies in More recently, low-income and middle- Lubet aging developed countries. The bottom- income countries such as South Africa, heavy population pyramids of Africa and China, and India have also begun their own the top-heavy pyramids of Europe ½t to- hrs surveys, which will provide a valuable gether hand in glove. However, integrat- look at population aging in emerging econ - ing the two over the next twenty-½ve years omies, for which data of this type are sore- would require immigrant flows from Af- ly lacking. rica to Europe that are more than ten times higher than current levels.25 Unfortunate- Counteracting the potential negative ly, most countries have gone in the oppo- con sequences of population aging will in - site direction and have instituted barriers volve some combination of behavioral and to immigration, usually in an effort to pro- policy changes. These may include in - tect their economies from low-wage work - creased rates of savings during the workers, to preserve traditions, to maintain cul - ing years; increased labor supply from tural and ethnic homogeneity, or to re- women, older people, and immigrants; spond to anti-immigrant sentiments. thoughtful policy programs; increased re- Meanwhile, Japan, another of the world’s tirement ages; and other adjustments. The most rapidly aging countries, has also had combination of interventions chosen will dif½culty attracting even highly skilled mi - determine how costs are divided among grant labor, in part due to language and current and future generations of older cultural factors and corporate promotion peo ple. Countries can ensure a smoother and pension systems designed around life- transition to an older population by initi- time, rather than medium-to-long term, ating policy and institutional reforms em ployment.26 It will be necessary to re- sooner. evaluate these priorities in order to pro- The solutions explored above may not duce migration policies that can meet both only ameliorate the potential economic the employment demand of young work- bur den of population aging, but may also ers from developing countries and the enhance the well-being of older people, care needs of older people in developed which is an independently valuable result. countries. Increased longevity can also be of great Encouraging research on population bene½t to society. Older people are repos- aging will lead to long-term solutions; itories of work experience, knowledge, and there are still many unanswered questions culture. Regarding them as a resource is in this ½eld. A great source of data for cur- crit ical to cultivating a more effective rent and future research lies in the family workforce and offsetting labor shortages. of Health and Retirement Studies (hrs). In a globalized world where knowledge Beginning with the ½rst hrs in the United and human capital provide an advantage, States, these longitudinal studies aim to the experience of older workers is valuable. collect rich and detailed data sets on older Positive perceptions of aging individu- adults and their families, covering every- als, their overall integration with society, thing from physical and mental health to and mitigation of ageist beliefs will also economic status and life histories. Coun- bene½t their quality of life. Indeed, sub- tries with the most rapidly aging popula- jective well-being–life satisfaction, feel-

144 (2) Spring 2015 89 Global ings of happiness, and sense of purpose shown that in recent decades the incidence Population and meaning in life–has been shown to be of disability at older ages is declining.28 Aging: Facts, closely related to physical health, a link Policy interventions can reinforce this Challenges, that is particularly important at older ages growing trend: higher socioeconomic sta- Solutions & 27 Perspectives and is associated with longer survival. tus and higher levels of education among Shifting the perspective away from an older men and women across multiple ageist view of burden and diminished val- countries correlates with reduced disabil - ue and toward a positive view of older peo - ity and chronic disease incidence.29 High- ple who offer experience, wisdom, and er education levels and improved health at leadership will enhance their lives and lead older ages has led to such sayings as “sev- to behaviors and policies with the power enty is the new sixty,” a sentiment rein- to resolve or offset the challenges present - forced by ½ndings that the average self- ed by aging. reported health of a sixty-nine-year-old Adopting policies that allow for healthy man in the United States in the 2000s was living throughout life and into old age will the same as that of a sixty-year-old-man encourage this more positive perspective in the 1970s. to take hold. Population aging has vastly Although demographic change is daunt- different potential for bene½t or harm to ing and has historically introduced signi½- societies depending on whether aging pop- cant obstacles to societal cohesion and eco- ulations are independent and healthy or nomic growth, the bottom line is that de- are disabled and in need of costly long- mography is not destiny. Individuals, bus - term care. As discussed above, it remains inesses, and governments have the option to be seen whether increased longevity will to adapt in the face of change, and many introduce improved quality of life and a solutions are within reach. The sooner pol- higher proportion of healthy years. While icy changes are considered and implemen- disability traditionally does increase along ted, the sooner population aging can trans- with age, research in the United States has form from a challenge into an opportunity.

endnotes * Contributor Biographies: DAVID E. BLOOM, a Fellow of the American Academy since 2005, is the Clarence James Gamble Professor of Economics and Demography in the Department of Global Health and Population at the Harvard T.H. Chan School of Public Health. His many publications include recent articles in such journals as JAMA: The Journal of the American Medical Association, Finance & Development, and Science. DAVID CANNING is the Richard Saltonstall Professor of Population Sciences and Professor of Economics and International Health in the Department of Global Health and Population at the Harvard T.H. Chan School of Public Health. His many publications include recent ar - ticles in such journals as Journal of Political Economy, Journal of Applied Statistics, and Journal of International Development. ALYSSA LUBET is a Research Assistant in the Department of Global Health and Population at the Harvard T.H. Chan School of Public Health. Her research interests include economics, women’s health and reproductive health, and population studies. Authors’ Note: Preparation of this manuscript was supported by Harvard University’s Program on the Global Demography of Aging, which receives funding from the National Institute on Aging, Grant No. 1 P30 AG024409-09.

90 Dædalus, the Journal ofthe American Academy of Arts & Sciences 1 Unless otherwise stated, population ½gures are drawn from United Nations Department of David E. Economic and Social Affairs Population Division, Population Estimates and Projections Bloom, Section, World Population Prospects: The 2012 Revision (New York: United Nations Depart- David ment of Economic and Social Affairs, 2014), http://esa.un.org/unpd/wpp/index.htm. Canning & Alyssa 2 Colin D. Mathers, Gretchen A. Stevens, Ties Boerma, Richard A. White, and Martin I. Tobias, Lubet “Causes of International Increases in Older Age Life Expectancy,” The Lancet 6736 (14) (2014). 3 Martin J. Prince, Fan Wu, Yanfei Guo, Luis M. Gutierrez Robledo, Martin O’Donnell, Richard Sullivan, and Salim Yusuf, “The Burden of Disease in Older People and Implications for Health Policy and Practice,” The Lancet 6736 (14) (2014): 100–111. 4 Sube Banerjee, “Multimorbidity–Older Adults Need Health Care that Can Count Past One,” The Lancet 6736 (14) (2014). 5 David E. Bloom, Elizabeth T. Ca½ero-Fonseca, Mark E. McGovern, Klaus Prettner, Anderson Stanciole, Jonathan Weiss, Samuel Bakkila, and Larry Rosenberg, “The Macroeconomic Impact of Non-Communicable Diseases in China and India: Estimates, Projections, and Comparisons,” The Journal of the Economics of Ageing 4 (2014): 100–111. 6 James F. Fries, “The Compression of Morbidity,” The Milbank Memorial Fund Quarterly 61 (3) (1983): 397–419, reprinted under the same title in The Milbank Quarterly (83) (4) (2005): 801–823. 7 Stacy L. Andersen, Paola Sebastiani, Daniel A. Dworkis, Lori Feldman, and Thomas T. Perls, “Health Span Approximates Life Span Among Many Supercentenarians: Compression of Morbidity at the Approximate Limit of Life Span,” The Journals of Gerontology: Biological Sci- ences & Medical Sciences 67 (4) (2012): 395–405; James F. Fries, Bonnie Bruce, and Eliza Chakravarty, “Compression of Morbidity 1980–2011: A Focused Review of Paradigms and Progress,” Journal of Aging Research 2011 (2011); and Riccardo E. Marioni, Michael J. Valenzuela, Ardo van den Hout, Carol Brayne, Fiona E. Matthews, and the mrc Cognitive Function and Ageing Study, “Active Cognitive Lifestyle is Associated with Positive Cognitive Health Tran- sitions and Compression of Morbidity from Age Sixty-Five,” PLOS One 7 (12) (2012): e50940. 8 Eileen M. Crimmins and Hiram Beltrán-Sánchez, “Mortality and Morbidity Trends: Is There Compression of Morbidity?” The Journals of Gerontology: Psychological Sciences & Social Sciences 66 (1) (2010): 75–86. 9 Somnath Chatterji, Julie Byles, David Cutler, Teresa Seeman, and Emese Verdes, “Health, Func - tioning, and Disability in Older Adults–Present Status and Future Implications,” The Lancet 6736 (14) (2014). 10 David E. Bloom, Axel Boersch-Supan, Patrick McGee, and Atsushi Seike, Population Aging: Facts, Challenges and Responses, Program on the Global Demography of Aging Working Paper No. 71 (Cambridge, Mass.: Harvard School of Public Health, 2012), http://www.hsph.harvard .edu/pgda/working.htm. 11 International Labour Organization, ILOSTAT Database (Geneva: International Labour Orga- nization, 2014), http://www.ilo.org/ilostat. 12 Benedict J. Clements, David Coady, Frank Eich, Sanjeev Gupta, Alvar Kangur, Baoping Shang, and Mauricio Soto, The Challenge of Public Pension Reform in Advanced and Emerging Market Economies, International Monetary Fund Occasional Paper 275 (Washington, D.C.: Interna- tional Monetary Fund, 2012). 13 David E. Bloom, David Canning, and Günther Fink, “Implications of Population Ageing for Economic Growth” Oxford Review of Economic Policy 26 (4) (2010): 583–612. 14 The World Bank, “World Bank Open Data,” http://data.worldbank.org (accessed January 26, 2015). 15 David E. Bloom, Somnath Chatterji, Paul Kowal, Peter Lloyd-Sherlock, Martin McKee, Bernd Rechel, Larry Rosenberg, and James P. Smith, “Macroeconomic Implications of Population Ageing and Selected Policy Responses,” The Lancet 6736 (14) (2014): 1–9.

144 (2) Spring 2015 91 Global 16 Gary S. Becker and Nigel Tomes, “Child Endowments and the Quantity and Quality of Chil- Population dren,” Journal of Political Economy 84 (4, pt. 2) (1976): S143–S162. Aging: 17 Facts, Klaus Prettner, David E. Bloom, and Holger Strulik, “Declining Fertility and Economic Well- Challenges, Being: Do Education and Health Ride to the Rescue?” Labour Economics 22 (C) (2013): 70–79. Solutions & 18 Perspectives Henry Aaron and Gary Burtless, eds., Closing the De½cit: How Much Can Later Retirement Help? (Washington, D.C.: The Brookings Institution, 2013). 19 David Bloom and David Canning, “How Companies Must Adapt for an Aging Workforce,” Harvard Business Review, December 3, 2012, https://hbr.org/2012/12/how-companies-must -adapt-for-a/. 20 The World Bank, “World Bank Open Data.” 21 Organisation for Economic Co-operation and Development, “Linking Pensions to Life Ex - pectan cy,” in Pensions at a Glance 2011: Retirement-Income Systems in OECD and G20 Countries (Paris: oecd Publishing, 2011), http://www.oecd-ilibrary.org/docserver/download/8111011e .pdf?expires=1422285516&id=id&accname=guest&checksum=CB440CD4A7F88C0E9274E171 B76C99BD. 22 Clements et al., The Challenge of Public Pension Reform in Advanced and Emerging Market Econ - omies; and Ronald Lee and Andrew Mason, “Fertility, Human Capital, and Economic Growth over the Demographic Transition,” European Journal of Population 26 (2) (2010): 159–182. 23 The United Nations, Political Declaration and Madrid International Plan of Action on Ageing, Second World Assembly on Ageing, Madrid, Spain, April 8–12, 2002. 24 David E. Bloom and Karen N. Eggleston, “The Economic Implications of Population Ageing in China and India: Introduction to the Special Issue,” The Journal of the Economics of Ageing 4 (2014): 1–7. 25 David Bloom, David Canning, and Jaypee Sevilla, The Demographic Dividend: A New Perspective on the Economic Consequences of Population Change (Santa Monica, Calif.: rand Corporation, 2002), http://www.rand.org/pubs/monograph_reports/MR1274.html. 26 Nana Oishi, “The Limits of Immigration Policies: The Challenges of Highly Skilled Migration in Japan,” American Behavioral Scientist 56 (8) (2012): 1080–1100. 27 Andrew Steptoe, Angus Deaton, and Arthur A. Stone, “Subjective Wellbeing, Health, and Ageing,” The Lancet 6736 (13) (2014). 28 National Research Council of the National Academies, Committee on the Long-Run Macro- economic Effects of the Aging U.S. Population, Aging and the Macroeconomy: Long-Term Impli - cations of an Older Population (Washington, D.C.: The National Academies Press, 2012). 29 Ahmad Reza Hosseinpoor, Jennifer Stewart Williams, Ben Jann, Paul Kowal, Alana Of½cer, Aleksandra Posarac, and Somnath Chatterji, “Social Determinants of Sex Differences in Dis- ability Among Older Adults: A Multi-Country Decomposition Analysis Using the World Health Survey,” International Journal for Equity in Health 11 (52) (2012).

92 Dædalus, the Journal ofthe American Academy of Arts & Sciences Individual & Social Strategies to Mitigate the Risks & Expand Opportunities of an Aging America

Julie M. Zissimopoulos, Dana P. Goldman, S. Jay Olshansky, John Rother & John W. Rowe

Abstract: Increasing life expectancy offers the potential bene½t of additional years of productivity and engagement to both individuals and society as a whole. However, it also carries substantial risks. For many, advanced age brings increased disease and disability (including cognitive impairment), ½nancial insecurity, and social isolation. These risks are greatest for those with the least education and ½nancial resources. An aging society must cope with increasing demands for high-quality geriatric care, mounting stresses on so - cial insurance programs (such as Social Security and Medicare in the United States), and the increasing danger that the growing gap between the haves and have-nots will threaten societal cohesion. These risks can be mitigated or aggravated by the lifestyle and savings behavior of individuals, families, employers, and the government. We present policy options in the areas of education, work and retirement, ½nancial security, health care, and social cohesion that can promote the bene½ts and reduce the risks of longer life.

JULIE M. ZISSIMOPOULOS is an The aging of America presents both opportunities Assistant Professor at the Sol Price and risks. The opportunities, which are often ne - School of Public Policy at the Uni - 1 versity of Southern California. glected, relate principally to the availability of large numbers of generally ½t, experienced older persons DANA P GOLDMAN . is Professor who can make valuable economic and social contri - of Public Policy, Pharmacy, and 2 Economics at the University of butions to society. Although increasing attention Southern California. is being paid to the risks of an aging society and ways to protect against them, the United States has yet to S. JAY OLSHANSKY is Professor of Epidemiology at the School of actively adapt to either the risks or the opportunities. Public Health at the University of The focus of this essay is to illuminate the changing Illinois at Chicago. nature of both. JOHN ROTHER is President and For individuals, the risks of a longer life include be- ceo of the National Coalition on coming ill or disabled and needing extensive acute Health Care. and long-term care. The latter is particularly impor- JOHN W. ROWE, a Fellow of the tant, as long-term care can deplete the patient’s ½ - American Academy since 2005, is nancial resources, leaving him or her dependent on Professor at the Columbia Univer - public support. Other individual risks include isola - sity Mailman School of Public Health. tion and disengagement: increasing num bers of older persons are living alone, with fraying connec- (*See endnotes for complete contributor biographies.) © 2015 by the American Academy of Arts & Sciences doi:10.1162/DAED_a_00333 93 Mitigating tions to families. U.S. families themselves smoking and greater educational gains Risks & are undergoing major changes in struc- may have bene½cial effects. The decision Expanding Opportu - ture and function that im pede their capac- to leave the labor force is not solely driven nities of an ity to serve as a safety net for elderly mem - by poor health but also by the intrinsic ½ - Aging bers. nancial incentives (or disincentives) in America Societal challenges and individual risks pub lic and private pension systems. Those are clearly linked. For instance, society is who are able and who choose to work long- ill-prepared to meet the surging demand er have more time to accumulate private for high-quality geriatric care. The short- savings and maximize public bene½ts. falls are not only in funding but in devel- Second, gains in life expectancy have not oping and maintaining a quali½ed health- been distributed equally across society. Ed - care workforce. Likewise, society has been ucation is closely tied to the “longevity div- slow to adapt to the need for “aging friend- idend”4: life-expectancy disparities by edu - ly” environments that are safe and afford- cation are astonishingly large.5 In 2008, able and provide transportation and hous- white males with sixteen or more years of ing appropriate for older residents. As the education lived 14.2 years longer than black United States faces the challenge of cover- men with less than twelve years of educa- ing the medical expenses of a growing tion. The same gap for women was 10.3 number of insured aging individuals, some years. And these disparities between the of the costs are being transferred to educational extremes continue to widen: healthier insured populations in the form in 1990, the gap in life expectancy between of higher out-of-pocket costs. Similarly, the most and least educated white females delays or reductions in Social Security was 1.9 years; in 2014, it was 10.4 years. retirement bene½ts, imposed to protect Finally, the current environment in the the solvency of the trust fund, place more United States is not conducive to disease- pressure on the elderly and their families. prevention efforts, even though evidence Finally, ensuring societal cohesion–en- suggests more proactive risk-factor pre- couraging an equitable society and lim- vention would yield high returns.6 With iting disparities and tensions between average tenure in health plans sometimes groups–is a societal challenge on a grand as low as four years–and with the govern- scale. ment ½nancing care after age 65–employ- Public and private strategies to mitigate ers and health plans clearly underinvest in the individual and societal risks of aging disease prevention. Thus, a much stronger must consider several fundamental issues. effort is needed to prevent disease before First, longer and healthier lives extend the it occurs. length of time that many people can engage productively in society. For others, Both individuals and institutions share however, innovations in medical science the responsibility to protect the elderly designed to prolong life can also prolong against the dual risks of physical frailty and periods of disability, which can limit abil- outliving their resources. In general, how - ity to work and increase medical expenses. ever, the strength of the forces that miti- Among those aged seventy and older, rates gate risk is eroding just as the risks them- of functional impairment have declined selves are growing. since the early 1980s, although more Social Security bene½ts, pensions, and recently the trend has flattened.3 Obesity private savings are the main economic re- may be contributing to declining rates of sources that support living expenses in re - physical function, though lower rates of tirement. This “three-legged stool,” how-

94 Dædalus, the Journal ofthe American Academy of Arts & Sciences ever, generally only supports high-income ment in an Aging America: The Role of Julie M. households,7 with lower-income house- Vol unteerism.”10 Zis si mo poulos, Dana P. holds often missing one or more of these Although some risky health behaviors Gold man, sources of support. Although social pro- (such as smoking) have declined over time, S. Jay Olshansky, John Rother & grams such as Social Security and Med i - others (such as poor diet and minimal John W. Rowe care may substitute for private savings, So- exercise) have increased, leading to rising cial Security alone will only provide enough obesity rates. Between 1970 and 2000, the resources to keep a household marginally prevalence of overweight children and ado - above the poverty line. Thus, for many lescents tripled, and obesity in adults dou- house holds, savings and private pensions bled, affecting 33 percent of the popula- –possibly augmented by support from tion.11 Evidence is accumulating that sug- other family members–are necessary to gests that this steady, decades-long increase pro vide for adequate retirement consump - may be abating, as the rise in obesity and tion. U.S. savings rates, however, have de- overweight rates appears to have flattened clined during the last thirty years.8 Even during the past decade.12 However, this if individuals designate private savings for may mean that some populations are reach- retirement, unexpected life events can ing a saturation point in obesity levels. Re - have profound effects on economic securi- search also points to evidence that al- ty later in life.9 Job loss, the loss of a spouse though being overweight is less damag ing through death or divorce, or ill health can to older persons than to the young, the con- disrupt savings plans and leave individuals sequences of the decades-long rise in obe- permanently worse-off, unable to regain sity will be felt for many years to come.13 their economic and noneconomic posi- Individual efforts at self-protection tions. The consequences of these shocks against health-related problems (and their can also affect families and households and attendant ½nancial stresses) are important. be transmitted intergenerationally when But they are often not enough to mitigate children have access to limited economic the risks associated with longevity, and resources and social support. families and other institutions are need- Health shocks are often unexpected ed to protect against ½nancial insecurity and sometimes unavoidable. However, in- and the consequences of physical frailty. dividuals may take actions over their life- times to delay or avert declining health. The American family has long served as Preventive health behaviors such as main- a vital safety net for older persons. Among taining a healthy diet, exercising regularly, older Americans who report needing at and avoiding harmful activities such as least some help with activities of daily liv- smoking can support long-term health. ing (adls, or simple activities such as feed- Changes in lifestyle (smoking, exercise) ing, dressing, and grooming oneself ) or in- and treatment of risk factors (high blood strumental activities of daily living (iadls, pressure, high cholesterol) have decreased or more complex skills such as managing the incidence of some diseases of old age, ½nances and transportation), almost one- including stroke and, more recently, heart half receive help from a family member.14 attack and some forms of cancer. In addi- Caregiving from family members is even tion, volunteering may offer substantial more prevalent and intensive in response health promotion and disease prevention to cognitive diseases such as dementia and bene½ts, as discussed by Dawn Carr, Lin da Alzheimer’s disease, which typically strike Fried, and John Rowe in their contribution at older ages. We have estimated in our past to this volume, “Productivity and Engage - research that Alzheimer’s disease results in

144 (2) Spring 2015 95 Mitigating almost $30,000 worth of unpaid caregiving generations, but at the same time, it cre- Risks & annually. If family members were un avail- ates greater obligations for the support and Expanding Opportu - able to provide this care–which ranges care of elderly family members when they nities of an from running errands and accompanying become frail and incapacitated. Aging older family members to the doctor to In sum, an aging society has clearly cre- America bathing, toileting, and administering med - ated more risk for the individual just as ications–on an unpaid basis, the addition- the buffers against these risks have begun al cost would be more than the total Med- to erode. This makes it all the more im- icaid spending for these individuals. 15 perative that effective, well-designed pub- Family support is central to ameliorat- lic programs are created to help. ing the risks of aging. As described more fully in the essay by Frank Furstenberg, Social insurance programs in the United Caroline Sten Hartnett, Martin Kohli, and States (both means-tested and not) play Julie Zissimopoulos in their contribu tion an important role in protecting individuals to this volume, “The Future of Intergen- against ½nancial insecurity. The leading so- erational Relations in Aging Societies,” the cial insurance programs are the Social Se - primary type of support provided by the curity retirement program (oasi: the Old- younger to the older generation in ad- Age and Survivors Insurance Trust Fund), vanced economies is caregiving.16 Finan- the Social Security Disability Insurance cial assistance is more likely to flow from program (di), Unemployment Insurance, the older generation to children and grand- Workers’ Compensation, and Medicare. children. Current and ongoing changes in Social insurance programs are intended to the structure and function of the family insure in divi duals against the risks of un - may disrupt these intergenerational trans- employment, disability, and old-age ½nan- fers of ½nancial and non½nancial resources, cial inse curity and inability to work. Be- thus compromising the family’s collective cause of their large scale, they have a great- ability to hedge against risks. er impact on poverty than means-tested Decreases and other changes in marriage programs such as Temporary Assistance and childbearing diminish the likelihood for Needy Families. of either children or spouses supplying The oasi program has the greatest im - care to a disabled older adult. Rates of mar- pact on poverty among the elderly. Without riage have declined and marriage has be - counting oasi, the poverty rate among old - come a less central and stable institution. er Americans was 55 percent in 2004. How- Childbearing, robust one half-century ago, ever, the poverty rate falls to nearly zero has slowed. A larger number of couples with oasi bene½ts. The Disability Insur- in which both spouses work has required ance program reduces rates of poverty women and men to develop more com- among the disabled to nearly zero as well.17 plex routines of managing work and fam- Work by the MacArthur Foundation Re- ily roles. Adding to this burden, greater in- search Network on an Aging Society ½nds vestment by parents is required today in that public expenditures on Amer icans childrearing and parental support: societal aged sixty-½ve and older are project ed to expectations of more higher education, as rise from $1.2 trillion in 2010 to $4.4 trillion well as the dif½culties young adults face in 2050. Public expenditures on the elderly in entering the workforce, have extended and disabled, while extremely ef fective in the period of young adults’ dependency on ensuring ½nancial security at older ages, parents. The rise in the number of seniors are not without drawbacks. Some studies adds potential capacity to assist younger posit that means-tested aid programs dis-

96 Dædalus, the Journal ofthe American Academy of Arts & Sciences courage saving, although the empirical evi- Medicare payments to private insurers and Julie M. dence on this is mixed.18 The evidence on medical providers, may also be a contrib- Zis si mo poulos, Dana P. the effect of oasi on saving is also mixed. uting factor. This trend however, may be Gold man, Social insurance programs do not only temporary and population aging will drive S. Jay Olshansky, John Rother & exist to buffer individuals from risk and up costs in the future. Moreover, the retire - John W. Rowe assist them when they are in need; they are ment of baby-boomers, combined with also a reflection of society’s values and of declining fertility rates producing fewer the greater social and political context. We workers, has dampened economic growth. are entering an era in which people are ex - Without new sources of government rev- pected to take increased personal respon- enue, slower growth reduces the amount sibility for their health and ½nancial situ- of revenue available for the social insur- ations. This is driven in part by the view ance programs the elderly have come to that our social insurance system has be - rely on. What role should individuals, fam- come unaffordable (driven largely by our ilies, society, and government play in mit- aging population), which has in turn cost igating the risks that old age brings? The young people their sense of ½nancial se - choices made today and in the future will curity. A successful aging society in the rede½ne U.S. society going forward. United States will require that each age group develop a sense of shared sacri½ce The growing momentum of the funda- and bene½t. It will also require policies that mental restructuring of private pensions promote economic growth and job oppor- in the last two decades–from traditional tunities for all ages. Finally, it will need a de½ned-bene½t plans to de½ned-contribu- social insurance system that reflects shift- tion plans such as 401(k) plans–is a critical ing risks and changes to the traditional buf- factor in the aging of the United States.21 fers against them. In their essay “Resetting De½ned-bene½t pensions are usually based Social Security” in this issue of Dædalus, on age, ½nal salary, and job tenure; they S. Jay Olshansky, Dana Goldman, and John gen erally provide a monthly income once Rowe also discuss in detail possible ap - the employee is eligible for full bene½ts and proaches to modi½cations in eligibility and retires. About 60 percent of the pension timing of Social Security bene½ts.19 wealth of the oldest baby-boomers resides The U.S. government faces the combined in de½ned-bene½t pension plans.22 On the challenge of the future ½nancial shortfall other hand, de½ned-contribution pension of the Social Security trust fund and rising values are not directly tied to age and ten- medical expenses of Medicare and Medi- ure, and they increase at rates that depend caid bene½ciaries. The growth in health on market return and whether and to what care costs has for many years far exceed- extent the employer decides to match con- ed the growth of gross domestic product tributions. The growing prevalence of de- (gdp). Some good news may be on the ho - ½ned- contribution plans will increase rizon, however: since 2010, real per capita workers’ opportunities to supplement So- health care spending has grown at an es - cial Security income but shifts investment timated annual rate of just 1.3 percent.20 risk from the business sector to the house- The causes of the slowdown are not yet hold. fully understood. Health care prices–not The rising number of de½ned-contribu - just use–are lower, implying that the slow- tion plans that pay in lump-sum distribu- down may be due to something more than tions rather than annuities also places the the recent recession. The Affordable Care responsibility for ½nancial decisions more Act’s Medicare reforms, which reduce squarely in an individual’s hands. This too

144 (2) Spring 2015 97 Mitigating applies to the increasingly common prac- toward a personal-responsibility model. Risks & tice of paying lump sums from de½ned- For example, many individuals now have Expanding Opportu - bene ½t plans. Older individuals must de- “consumer-directed health plans” that car- nities of an cide whether to spend the payout imme- ry high deductibles and encourage indivi- Aging diately or whether to roll it over and save duals to control their use of health care ser- America it. If they decide to save it, they must deter - vices. Medicare bene½ciaries must now mine how to invest it and at what rate they choose from a large menu of insurance should spend it in retirement in order to choices when signing up for prescription not outlive their savings. In addition, the drug coverage through Medicare Part D. greater prevalence of de½ned-contribution Choosing an optimal health insurance plan, pension plans with more retirement assets however, requires an understanding of in - held in stocks have made retirement plans surance terms that many people are unfa- subject to changes in the stock market, as miliar with26 and an understanding of the recent recession has underscored. how different bene½t designs affect out-of- The movement from de½ned-bene½t to pock et spending.27 Many older adults do de½ned-contribution plans is one example not understand the unique design of Part of how individuals are being asked to man- D plans, a knowledge de½cit that exposes age their own ½nances and retirement as - them to a coverage gap.28 This gap may sets. The shift to a “personal-responsibil- have long-term health consequences when, ity” retirement model will only be success- as research shows, bene½ciaries with ful if ½nancial literacy rises. Financial in - chronic disease must forgo their medica- struments are becoming more complex, tions as a result.29 and some individuals may be ill-equipped to make complex investment decisions. Despite these challenges, feasible policy The fact that many elderly do not choose options can lower the overall risk to both the best Medicare Part D (drug bene½t) individuals and society and increase the plan is just one example of this de½cit.23 Fi - likelihood that the United States will re - nancial literacy is highest among the most main cohesive, productive, secure, and educated, but the rate of college com ple - equitable as it emerges from this demo- tion is flattening, so ½nancially under-lit- graphic transformation. Based on the in - erate populations will likely rise if this formation presented in this essay and in trend continues.24 the other essays in this volume, the Mac - Other buffers protecting ½nancial se curi - Arthur Foundation Research Network on ty in retirement are also eroding. Employers an Aging Society offers the following ½ve today are less likely to offer their work ers recommendations as a way forward. retiree health insurance bene½ts. Ac cord - ing to the Employee Bene½t Research Insti- 1) Enhance life-long learning and increase tute (ebri), in 2010, 17.7 percent of work- the likelihood that older workers can func- ers were employed in establishments that tion effectively in the labor force: offered health coverage to early retirees, • Offer incentives for reinvesting in skill down from 28.9 percent in 1997.25 Those development, especially for blue-collar that continue to offer retiree health bene - workers. ½ts have made changes in the bene½ts they • Encourage work site–based education- offer, including raising premiums, tight - al and training programs. ening eligibility, and reducing bene½ts. Just as with ½nancial models in retire- • Provide resources that support alterna- ment, there has been a shift in health care tive ways to update skills; encourage

98 Dædalus, the Journal ofthe American Academy of Arts & Sciences lifelong learning beyond the classroom. oth ers) with additional training and Julie M. An example is the Mozilla/MacArthur de monstrated competence in geriat - Zis si mo poulos, Dana P. Foundation Badges for Lifelong Learn- rics. This must include the diagnosis Gold man, ing Program, an emerging model of and management of common disor- S. Jay Olshansky, John Rother & peer-to-peer learning that creates cre - ders of late life, including delirium, John W. Rowe dentials for informal learning that is dementia, falls, incontinence, poly - not currently captured by traditional pharma cy (use of more than four med- credentials (such as college degrees). ications, which raises the possibility of harmful interactions), and frailty; as 2) Ensure that older persons are produc- well as diseases especially common in tively engaged in society, either through older persons, such as diabetes, cardio- paid work or volunteering: vascular disease, and arthritis. • Create incentives for employers to of - • Establish training programs tailored fer more flexible employment models to speci½c caregiver situations. that ½t the needs of older workers. • Launch caregiver support programs as • Create incentives for volunteering. a place to discuss challenges in a con- • Increase funding for federal senior vol- ½dential setting. unteer programs. • Expand the National Family Caregiver • Consider the bene½ts and the costs of Support Program to help reimburse establishing Medicare as the primary costs of caregiving and provide fund- payer for health bene½ts of older working for caregivers to be temporarily re- ers who are eligible for the program. lieved of their duties, allowing them to 3) Encourage individual and societal ½- maintain their responsibilities for an nancial security: extended time. • Change 401(k) participation from vol- 5) Build a culture of shared sacri½ce and untary opt-in to a default option of bene½t across the generations: par ticipation and a voluntary opt-out, • Promote generationally cohesive com - and require savings rates of 6 percent. munities by encouraging individuals • Offer paid leave for family caregiving. from different generations to interact • Ensure that Social Security continues productively together, reducing inter- to provide individuals and families generational tensions. with ½nancial security while shoring • Establish programs to involve seniors up trust fund solvency through bene½t in schools and youth in senior services. reform. • Avoid exclusive reliance on age-segre- 4) Provide high-quality health care to all: gated housing and services. • • Strengthen geriatric training and in - Emphasize the bene½ts of social insur - crease requirements for demonstrated ance programs across the entire age competence in geriatric care for all lev - span. els of health care providers. Taken together, these options provide a • Provide ½nancial incentives such as general blueprint for the types of policies loan forgiveness and scholarships for that should be put in place to increase the individuals training in geriatric care. likelihood that our country will maintain • Channel resources through Medicare its resilience in the face of this demograph - to providers (nurses, physicians, and ic change.

144 (2) Spring 2015 99 Mitigating endnotes Risks & JULIE M ZISSIMOPOULOS Expanding * Contributor Biographies: . is an Assistant Professor at the Sol Price Opportu - School of Public Policy and the Associate Director of the Schaeffer Center for Health Policy nities of an and Economics at the University of Southern California. Her research has been published in Aging such journals as The Journal of Economic Perspectives, Journal of Health Economics, Journal of America Human Resources, and Demography. DANA P. GOLDMAN is the Leonard D. Schaeffer Chair in Health Policy at the University of Southern California. He is also the Director of the Schaeffer Center for Health Policy and Economics. He serves as a health policy adviser to the Congressional Budget Of½ce, and his research has appeared in the New England Journal of Medicine, JAMA: The Journal of the American Medical Association, Demography, Journal of the American Statistical Association, and Health Affairs. S. JAY OLSHANSKY is Professor of Epidemiology at the School of Public Health, Division of Epidemiology and Biostatistics at the University of Illinois at Chicago. He has published articles in such journals as The New England Journal of Medicine, JAMA, Science, The Scientist, Scienti½c American, and Health Affairs. JOHN ROTHER is President and ceo of the National Coalition on Health Care. Prior to joining the Coalition, he served as the Executive Vice President for Policy, Strategy, and International Affairs at the aarp. He was also Staff Director and Chief Counsel of the Senate Special Com - mittee on Aging under the direction of Chairman John Heinz (r-pa) and Special Counsel for Labor and Health to Senator Jacob Javits (r-ny). JOHN W. ROWE, a Fellow of the American Academy since 2005, is Professor at the Columbia University Mailman School of Public Health and Chair of the MacArthur Foundation Research Network on an Aging Society. He is the author of Successful Aging (with Robert L. Kahn, 1998) and was the Chair of the Institute of Medicine of the National Academies project the Future Health Care Workforce for Older Americans, which authored the report Retooling for an Aging America: Building the Health Care Workforce (2008). 1 One exception is John Wallis Rowe and Robert L. Kahn, Successful Aging (New York: Dell, 1998). 2 David Lowsky, S. Jay Olshansky, Jay Bhattacharya, and Dana P. Goldman, “Heterogeneity in Healthy Aging,” The Journals of Gerontology: Biological Sciences & Medical Sciences 69 (6) (2014): 640–649. 3 Robert Schoeni, Vicki Freedman, and Linda Martin, “Why is Late-Life Disability Declining?” Milbank Quarterly 86 (1) (2008): 47–89. Among the working-age population, results are con- flicting, in part because of variation in how disability is de½ned and measured in surveys. See Richard Burkhauser, Andrew Houtenville, and Jennifer Tennant, “Measuring the Population with Disabilities for Policy Analysis” in Lifecycle Events and Their Consequences: Job Loss, Family Change and Decline in Health, ed. Kenneth A. Couch, Mary C. Daly, and Julie M. Zissimopoulos (Stanford, Calif.: Stanford University Press, 2013). 4 S. Jay Olshansky, Daniel Perry, Richard A. Miller, and Robert N. Butler, “In Pursuit of the Longe vity Dividend,” The Scientist 20 (3) (2006): 28–36. 5 S. Jay Olshansky, Toni Antonucci, Lisa Berkman, Robert H. Binstock, Axel Boersch-Supan, John T. Cacioppo, Bruce A. Carnes, Laura L. Carstensen, Linda P. Fried, Dana P. Goldman, James Jack son, Martin Kohli, John Rother, Yuhui Zheng, and John W. Rowe, “Differences in Life Ex - pectancy Due to Race and Educational Differences are Widening, and Many May Not Catch Up,” Health Affairs 31 (8) (2012): 1803–1813. 6 Dana P. Goldman, Yuhui Zheng, Federico Girosi, Pierre-Carl Michaud, S. Jay Olshansky, David Cutler, and John W. Rowe, “The Bene½ts of Risk Factor Prevention in Americans Aged 51 Years and Older,” American Journal of Public Health 99 (11) (2009): 2096–2101. 7 John Karl Scholz, Ananth Seshadri, and Surachai Khitatrakun, “Are Americans Saving ‘Opti- mally’ for Retirement?” Journal of Political Economy 114 (4) (2006): 607–643.

100 Dædalus, the Journal ofthe American Academy of Arts & Sciences 8 U.S. Department of Commerce, “Seasonally Adjusted Annual Rate,” psavert Series (Washing- Julie M. ton, D.C.: Department of Commerce, 2014). Zis si mo poulos, Dana P. 9 Kenneth Couch, Mary C. Daly, and Julie M. Zissimopoulos, eds., Lifecycle Events and Their Conse - Gold man, quences: Job Loss, Family Change and Decline in Health (Stanford, Calif.: Stanford University Press, S. Jay Olshansky, 2013). John Rother & John W. Rowe 10 Dawn C. Carr, Linda P. Fried, and John W. Rowe, “Productivity and Engagement in an Aging America: The Role of Volunteerism,” Dædalus 144 (2) (2015). 11 Katherine M. Flegal, Margaret D. Carroll, Cynthia L. Ogdent, and Lester R. Curtin, “Prevalence and Trends in Obesity among U.S. Adults, 1999–2008,” JAMA 303 (3) (2010): 235–241. 12 Katherine M. Flegal, Margaret D. Carroll, Brian K. Kit, and Cynthia L. Ogden, “Prevalence of Obesity and Trends in the Distribution of Body Mass Index among U.S. Adults, 1999–2010,” JAMA 307 (2012): 491–497. 13 Linda G. Martin, Vicki A. Freedman, Patricia M. Andreski, and Robert F. Schoeni, “Recent Trends in Disability and Related Chronic Conditions among People Ages 50 to 64,” Health Affairs 29 (4) (2010): 725–731. 14 These calculations are based on the 2004 survey of the Health and Retirement Study; see University of Michigan Health and Retirement Study (hrs), http://hrsonline.isr.umich.edu. 15 Lynn Feinberg, Susan C. Reinhard, Ari Houser, and Rita Coula, Valuing the Invaluable: 2011 Update–The Growing Contributions and Costs of Family Caregiving (Washington, D.C.: aarp Public Policy Institute, 2011), 2. 16 Frank F. Furstenberg, Caroline Sten Hartnett, Martin Kohli, and Julie M. Zissimopoulos, “The Future of Intergenerational Relations in Aging Societies,” Dædalus 144 (2) (2015). 17 Yonatan Ben-Shalom, Robert A. Mof½tt, and John Karl Scholz, “An Assessment of the Effec- tiveness of Anti-Poverty Programs in the United States,” nber Working Paper 17042 (Cam- bridge, Mass.: National Bureau of Economic Research, 2011). 18 See, for example, Erik Hurst and James P. Ziliak, “Do Welfare Asset Limits Affect Household Saving? Evidence from Welfare Reform,” Journal of Human Resources 41 (1) (2006): 46–71. 19 S. Jay Olshansky, Dana P. Goldman, and John W. Rowe, “Resetting Social Security,” Dædalus 144 (2) (2015). 20Executive Of½ce of the President of the United States, Council of Economic Advisers, Trends in Health Care Cost Growth and the Role of the Affordable Care Act (Washington, D.C.: Executive Of½ce of the President of the United States, November 2013), http://www.whitehouse.gov/ sites/default/½les/docs/healthcostreport_½nal_noembargo_v2.pdf. 21 Nicole Maestas and Julie Zissimopoulos, “How Longer Work Lives Ease the Crunch of Pop- ulation Aging,” Journal of Economic Perspectives 24 (1) (2010): 139–160. 22 Alan Gustman, Thomas Steinmeier, and Nahid Tabatabai, “What the Stock Market Decline Means for the Financial Security and Retirement of the Near-Retirement Population,” Journal of Economic Perspectives 24 (1) (2010): 161–182. 23 Florian Heiss, Adam Leive, Daniel McFadden, and Joachim Winter, “Plan Selection in Med - icare Part D: Evidence from Administrative Data,” Journal of Health Economics 32 (6) (2013): 1325–1344. 24 Organisation of Economic Co-operation and Development, Education at a Glance, 2011: OECD Indicators (Paris: oecd Publishing, 2013). 25 Paul Fronstin and Nevin Adams, “Employment-Based Retiree Health Bene½ts: Trends in Access and Coverage, 1997–2010,” EBRI Issue Brief (377) (October 2012). 26 Silvia Barcellos, Amelie C. Wuppermann, Katherine Grace Carman, Sebastian Bauhoff, Daniel L. McFadden, Arie Kapteyn, Joachim K. Winter, and Dana Goldman, “Preparedness of Amer -

144 (2) Spring 2015 101 Mitigating icans for the Affordable Care Act,” Proceedings of the National Academy of Sciences 111 (15) (2014): Risks & 5497–5502. Expanding 27 D Opportu - Heiss et al., “Plan Selection in Medicare Part .” nities of an 28 John Hsu, Vicki Fung, Mary Price, Jie Huang, Richard Brand, Rita Hui, Bruce Fireman, and Aging D America Jo seph P. Newhouse, “Medicare Bene½ciaries Knowledge of Part Prescription Drug Program Bene½ts and Responses to Drug Costs,” JAMA 299 (16) (2008): 1929–1936. 29 Geoffrey F. Joyce, Julie M. Zissimopoulos, and Dana P. Goldman, “Digesting the Doughnut Hole,” Journal of Health Economics 32 (6) (2013): 1345–1355.

102 Dædalus, the Journal ofthe American Academy of Arts & Sciences Board of Directors Don M. Randel, Chair of the Board Jonathan F. Fanton, President Diane P. Wood, Chair of the Council; Vice Chair of the Board Alan M. Dachs, Chair of the Trust; Vice Chair of the Board Jerrold Meinwald, Secretary Carl H. Pforzheimer III, Treasurer Nancy C. Andrews Helene L. Kaplan Ira Katznelson Nannerl O. Keohane Roger B. Myerson Venkatesh Narayanamurti Pauline Yu Louis W. Cabot, Chair Emeritus

Inside back cover: Edna Warf, a retired woman who volunteers as a literacy tutor at Claxton Elementary School in Asheville, North Carolina, reads alongside ½rst-grade student Valencia Barton. Volunteering in schools has been shown to produce health bene½ts for elderly volunteers while also addressing a signi½cant societal need. © Owen Franken/Corbis Images.

Dædalus coming up in Dædalus:

Food, Health & G. David Tilman, Walter C. Willett, Meir J. Stampfer & Jaquelyn L.

U.S. $14; www.amacad.org Cherishing Knowledge · Shaping the Future