Late-Life Inequality in the Second Gilded Age: Policy Choices in a New Context

Over the decades, the topic of income inequality has had its ups and downs in public policy discourse in the United States. During the Great Society period of the Kennedy and Johnson Administrations, the plight of those lowest in the income distribution was a focus of attention under the rubric of the “war on poverty.” From the 1980s into the early years of the twenty-first century, discussion of inequality was to a significant extent displaced by themes of economic growth and the burdens of taxation, with explicitly redistributional benefit and taxation policies losing favor. In some quarters, attention to issues of inequality was criticized as a distraction, or worse; for example, in 2004, an influential Chicago School economist wrote, “Of the tendencies that are harmful to sound economics, the most seductive, and in my opinion the most poisonous, is to focus on questions of distribution.” ( Lucas, 2004 ). During this period, the progressivity of the income tax system was reduced through reductions in top marginal tax rates and capital gains taxation, and there was a significant shift from traditional defined-benefit to defined-contribution retirement plans in the retirement income system. During this period, as well, the cost of entitlements became increasingly contested, with the “generational equity” narrative increasingly promoted in support of policy changes to reduce these costs. As Beard and Williamson note in this issue of the Public Policy and Aging Report ( PPAR ), this narrative has been particularly prominent in calls to reduce the benefits payable under current law to the cohorts of individuals who will reach late life in coming decades—calls that, while historically more identified with Republican legislators, have received considerable support among national elected officials on both sides of the aisle. Even President Obama, in his 2014 budget proposal, included a proposal to shift future cost-of-living increases for Social Security from the standard Consumer Price Index to a “chained CPI” formula that would reduce benefits by approximately 2% over the course of individuals’ retirement ( Calmes, 2013 ; Van de Water & Ruffing, 2013 ), although this proposal included some protections for beneficiaries with lower benefit levels and was not included in his subsequent budget proposals. More substantial, recurring proposals call for increasing the age of eligibility for full Social Security benefits, which would have the effect, for each year of postponed eligibility, of a roughly 7% across-the-board cut in benefits, regardless of whether a worker files for Social Security before, upon, or after reaching the full retirement age ( Van de Water & Ruffing, 2013 ). Within gerontology and social science, more generally, attention to economic inequality over the life course, including late life, has also had its ups and downs. As noted by Beard and Williamson in this issue of PPAR , long-standing assumptions in gerontology asserted that economic inequality narrows as we age, with benefit programs replacing the labor market as income sources, but this assumption was contradicted by analyses by Crystal and Shea (1990a) , who concluded that a process of snowballing economic advantage and disadvantage over the life course is instead in play. Initially, this process of diverging outcomes over the life course, leading to highly disparate late-life circumstances, was described as the “two worlds of aging” phenomenon by Crystal (1982 , 1986 ); a related concept was described as the phenomenon of “intracohort differentiation” by Dannefer (1987) . In a line of publications beginning with their 1990 paper on “Cumulative Advantage, Cumulative Disadvantage, and Inequality Among Elderly People,” Crystal and Shea (1990a , 1990b ) introduced into the gerontological and social science literature the “cumulative advantage” term to describe this process, in a theoretical model that described and examined the process of production of disparate life outcomes as one of iterative interaction of initial advantages and societal institutions over the life course ( Crystal, 2006 ; Crystal & Shea, 1990a , 1990b ; Crystal, Shea, & Krishnaswami, 1992 ; Crystal & Waehrer, 1996 ). The “cumulative advantage” concept and terminology came to be widely adopted from the 1990s forward (e.g., Dannefer, 2003 ; O’Rand, 1996 ), although much of the attention to cumulative advantage processes focused on cumulation of health disparities rather than those in income. Nevertheless, the issue of income inequality has been part of the cumulative advantage tradition in gerontology since its inception. Recently, Crystal, Shea, and Reyes (2016) have updated their earlier work on late-life inequality, finding further increases in inequality in recent years. Between 1983–84 and 2010, the share of total income received by 65-to-74 year olds in the lower 40% of the income distribution went down from 17% to 14%, while the share of the best-off 20% increased from 46% to 48%. For people 75 and older, the share received by the lower 40% decreased from 15% to 14%, while the share of the top 20% increased from 47% to 50%, with the lower 80% receiving only 50%. Concentration of wealth among older people also increased. By 2010, the top 20% accounted for 62% of total annuitized value of wealth. Overall, inequality within each birth cohort increased as its members moved from early to mid to later life, both for the cohorts that make up the current elderly and for the very high-inequality cohorts now in midlife. In broader public discourse related to inequality, although there was limited focus on the issue in the last decades of the twentieth century, the topic has gained renewed attention since the start of the twenty-first century, as evidence has accumulated that the nation has entered a new, higher inequality era. The unexpected rise of Thomas Piketty’s Capital in the Twenty-First Century ( Piketty, 2014 ) to the top of national bestseller lists was one indicator of increasing concern with this growing problem, which also emerged in 2015 as a central focus of Bernard Sanders’s campaign for the Democratic nomination for the presidency. Utilizing income tax records that capture very high incomes more fully than is typically the case with survey data on income, Piketty and his colleague Emanuel Saez were able to highlight the increasing concentration of income in the very top of the income distribution, especially the top 1%, in a way that has struck a responsive chord that has resonated with rising concerns about the deleterious social effects of growing income inequality. As Piketty and Saez document, this trend is of international scope, although particularly marked in the United States ( Piketty & Saez, 2014 ). In contrast with the dominant themes of the War on Poverty era in the United States, which concentrated on the plight of the poorest, the more recent wave of concern with income inequality has highlighted concerns with broader measures of income inequality and, in particular, the growing economic and social gap between the wealthiest Americans, whose income has sharply increased in recent decades, and the circumstances of those with incomes that are broadly in the lower part of the income distribution, who have experienced income stagnation. Although the growth in income inequality has affected all age groups, the increased public attention to the issue has focused predominantly on the nonelderly population, with less attention thus far to implications for late-life inequality, the way in which it has been affected by economic and social changes, and implications for future retirement income policy. However, these implications are substantial and have direct bearing on the consequences for equity of key policy choices, including the financing and benefit levels of Social Security, the shift from defined-benefit to defined-contribution retirement plans ( Johnson & Crystal, 2003 ), the role of substantial tax expenditures in supporting these plans, and many other policy choices. It is now apparent that we are now in a period of higher inequality than has been seen for almost a century—both in the United States and in other developed countries—although the United States is unfortunately in the “lead” ( Piketty and Saez, 2014 ). This development creates a new context for retirement income policy in the twenty-first century as well as for other policies affecting late life, such as the provision and financing of health and long-term care programs. In hindsight, it appears that the four decades following World War II were a period in which the benefits of economic growth were unusually broadly distributed compared with both earlier and later eras. For example, work by Piketty (2014) and Saez (2013) has shown that across multiple measures of income, the income share of the bottom 90% was quite stable during the post–World War II period up to the early 1980s and markedly greater than the period prior to 1940. Beginning in the 1980s and continuing to the present, however, there has been a steady increase in the level of overall inequality, with the income share of the bottom 90% of the population across all ages estimated to have declined from the prevailing level of the postwar period (approximately 65%) to about 50%. After a slight dip in 2007 through 2009 during the recession, the trend to increased disparity resumed during the 2009–2012 period. By 2012, the income share of the top 10%, at 50.4%, slightly exceeded its previous peak in 1928, the peak of the stock market bubble in the Gilded Age of “roaring” 1920s ( Saez, 2013 ). The United States appears to have settled into a new economic era of higher inequality with lowering of trade barriers, reduced demand for less-educated lower skilled workers, decline of labor unionism and well-paid industrial jobs with few educational requirements, growing educational debt, reduced income, and inheritance taxes on higher income individuals among a host of contributing and converging factors. With levels of inequality now comparable to those of the original Gilded Age, the current era could reasonably be viewed as the Second Gilded Age. This new “Gilded Age” high-inequality environment raises new concerns about the potential impact of proposed policy changes aimed at limiting the cost of “entitlements” for individuals in the lower part of the income distribution, among the current elderly persons and, perhaps even more seriously, among cohorts who will be reaching late life in the coming decades. As Johnson notes in this issue, members of these cohorts have been able, to some degree, to buffer the impact of wage stagnation through delays in retirement age (reversing trends of earlier years) and through increased labor force participation by women, resulting in higher proportions of families having multiple earners. However, there are natural limits to the ability of these behavioral changes to offset wage stagnation, and lower income workers have experienced increasing difficulty in accruing meaningful retirement savings, particularly for less-educated workers whose prospects have significantly declined. The projections from the Urban Institute’s DYNASIM model, presented by Richard Johnson in this issue, provide a sobering picture of the distributional outlook for the cohorts that will be reaching age 70 through the early 2040s. Although, as Johnson notes, overall improvement is projected to rise in inflation-adjusted terms for these cohorts, income inequality is projected to increase for these cohorts, consistent with the general pattern of broadly increasing inequality documented by Piketty and others. Given the increased concentration of income and wealth in the very top of the income distribution, and the well-documented undercapture of income and wealth in this part of the distribution in survey data ( Crystal, 1986 ; Piketty, 2014 ; Smith, 2003 ), the DYNASIM estimates of the magnitude of increasing late-life inequality are likely conservative. Nevertheless, as Johnson notes, the projections indicate a very high-inequality retirement income future, with a seven-to-one disparity projected at age 70 between those at the 90th percentile of income and those at the 25th percentile, among those born between 1970 and 1974. Of course, although more difficult to project with precision, disparities between those in the top 5% or top 1% and lower income cohort members will be much greater. The increasing gap in circumstances and lifestyle between those at the top and the rest of the older population may have broad policy consequences, given the considerable influence of the most-advantaged group. For example, with ample private resources, the most advantaged may have declining stakes in publicly funded programs that support long-term care, comprehensive coverage of medical services, and financial transfers such as Social Security retirement pensions. Another concern about prospects for lower socioeconomic status (SES) elderly in coming years involves disparities in health status and other indicators of well-being. As Johnson notes, it is well known that health problems are more common among lower income than higher income adults, and this disparity extends to the elderly as well. These disparities will limit the ability of lower SES individuals to supplement their income with employment and contribute to their out-of-pocket cost burden ( Crystal, Harman, Sambamoorthi, Johnson, & Kumar, 2000 ; Johnson, 2013 ). Given the disparity in health status, disparities in out-of-pocket cost burden could increase if widely discussed proposals to limit the growing taxpayer costs of Medicare by shifting some costs to beneficiaries and increasing their “skin in the game” are adopted, along the lines of current trends in insurance designs for the nonelderly. Proposals to increase consumer choice in Medicare benefit design, by tiering plans into more comprehensive coverage at higher cost to beneficiaries and limited coverage at lower cost, would likely leave lower SES individuals with less protection against out-of-pocket costs than more advantaged individuals. Although the health–wealth association in late life is well known, less well known but sobering data from recent studies suggest that gaps in various indicators of well-being may be growing among cohorts currently at midlife who will constitute the elderly of coming decades. For example, Phillips, Robin, Nugent, and Idler (2010) found that suicide rates for less-educated men and less-educated women increased markedly between 2000 and 2005 among 40–49 year olds and especially among 50–59 year olds. Suicide rates increased by 27% for men aged 50–59 with high school or less education while remaining essentially flat for college-educated men. Educational disparities in suicide rates increased; in 2005, the suicide rate for men with a high school education or less was about 2.0 times the rate among those with some college education and 2.6 times that of those with at least a college degree, with disparities by education almost as great among women. In recent work, Case and Deaton (2015) reported rising rates of both morbidity and mortality in midlife (ages 45–54) among white non-Hispanic Americans in the twenty-first century. Their analysis of National Center for Health Statistics data (National Health Interview Surveys, the Behavioral Risk Factor Surveillance System, and the National Health and Nutrition Examination Survey) reported declines between 1997–1999 and 2011–2013 in self-rated health, days physical health was “not good,” pain, mental health (Kessler-6 score), and measures of functioning in activities of daily living as well as obesity and heavy drinking. These adverse trends are likely most severe among the lower SES members of these cohorts. Cumulative advantage theory would predict that the adverse trends observed in these cohorts at midlife would carry forward into the retirement years, further increasing concern about retirement-age well-being among less-advantaged members of the cohorts reaching late life in coming decades. These are the same cohorts that are the targets of widely touted proposals to reduce the cost of future entitlements by increasing the age of eligibility for full Social Security benefits, reducing cost-of-living adjustments scheduled under current law, or incorporating more defined-contribution elements into Medicare, among other changes. Examining data from the 1970s, Crystal (1982) described the policy challenges posed by a pattern of late-life inequality characterized as the “two worlds of aging” phenomenon. Subsequent work in the cumulative advantage tradition has highlighted the processes by which early advantages and disadvantages have enduring effects on a range of outcomes into late life. This now-robust body of research demonstrates, among other the of educational on outcomes the life that is likely to be even more for coming of elderly of the broad societal and economic changes that have increased financial to and reduced life for those with lower levels of a and a into the twenty-first century, we at a the effects of broad economic changes leading to greater economic inequality across the age documented by Piketty (2014) over an enduring pattern of cumulative advantage and disadvantage over the life course that has now been documented to be an within multiple cohorts over their life course ( Crystal & Shea, 2003 ; Crystal & Waehrer, 1996 ). The effects of these a and, an of the “two worlds of aging” for coming cohorts of These highlight the of attention to policy that, to an buffer the both of economic trends and among in the United States, is the Social Security which an economic that is especially for lower income for more than of the income of elderly individuals in both the lowest and the of income Shea, & In the United States, other has emerged that its particularly for those of who only limited benefit from retirement and are to meaningful retirement savings, even are In of the of rising inequality across the age and the of cumulative advantage and disadvantage processes in late-life there is to for rather than reducing future Social Security benefits payable under current For example, projected in the Social Security could be by or the on the of income to the Social Security Given the lower incomes and greater expenditures for that are of over age a benefit increase also be for the to a provision that was as a into President 2013 proposal to reduce cost-of-living adjustments the “chained consumer ( Van de Water & Ruffing, 2013 ). aimed at the elderly who are or receiving the lowest Social Security benefits, including the Security could also be at costs that are in the overall of retirement changes could be by adjustments to the estimated in current tax that currently retirement private and other retirement income that benefit higher income individuals. Although such changes are difficult in an in which tax as a of has declined in the twenty-first century the of the population ( 2014 ), the may not be the of given of increasing public about the consequences of growing income inequality. on the of current benefit Social Security highly In the a the for less-advantaged members of cohorts of elderly of the protections by currently scheduled Social Security benefit Although growing income inequality is an international as documented by Piketty, policy on the and financing of retirement income institutions in or not the impact of these and of cumulative advantage processes that contribute to high late-life inequality. As (2013) has is a . . . Of the has some of the disparities in incomes and with consequences . . . inequality . . . has as we have in education and health care and social inequality by system and . . . For these a not between the and but also between those that about and those that will be in . . . will inequality for the of policies in the future in significant part on of policies are able to the issues of retirement income and other policies in the For example, as Beard and Williamson note in this issue, the “generational equity” of about the cost of entitlements has to benefit as on the of the equity by equity by of between age groups, the disparities within age however, the of equity most to equity among rather than between age attention to the of economic inequality across individuals some for that the of increasingly disparate and current distributional to these may data contribute to the of the elderly as a population ( & ). are at a disadvantage to in such but have a role to as of in a to publicly to social problems is will be work to of inequality, to the undercapture of financial inequality in survey data and to further if evidence of on multiple among less-advantaged members of the cohorts that will constitute the elderly population in coming decades. is a continuing for and more widely data and analyses that to focus about retirement benefits on equity among as issues of in of “generational equity” across highly disparate analyses of the current and future of inequality have an role to in an policy about the continuing of retirement income in a economic and policy in which the challenges of cumulative advantage and disadvantage within individuals’ life with the new of increasing overall income inequality.