In the coming decades the United States will undergo a demographic change as important as any in its history. This change will have significant economic and fiscal effects, many of which have been investigated in the preceding pages. This report confirms what other analyses have posited with regard to the nation’s fiscal situation: The fiscal effects of population aging will be very large and will be mediated by how and how quickly our society responds. With regard to macroeconomic effects, the report concludes that while the overall macroeconomic effects of aging may be modest for the economy as a whole, the risks could be large for particular age groups and generations if the burden of adjustment is borne by public programs such as Social Security, Medicare, and Medicaid or if the costs are borne largely by subgroups such as workers or retirees.
Throughout its deliberations, the committee was aware that trends in some of the topics under discussion were, by their nature, speculative. There was a tension between suggesting potential pathways of future change and, at the same time, remaining grounded in empirical knowledge. There were many questions concerning the macroeconomic effects of population aging that the committee thought were important but could not address fully because of a lack of data and/or research. Part of the committee’s charge, in response to a request from the Division of Behavioral and Social Research, U.S. National Institute on Aging (NIA), was to identify and recommend major research needs related to the macroeconomics of aging.
This chapter presents the committee’s thinking about which additional analyses and lines of inquiry would be most useful, taking into account the feasibility of research on a given topic. Topics have been grouped into four
categories: (1) demographic and health measurement and projections; (2) capacity to work and longer working life; (3) changes in consumption and saving; and (4) modeling efforts and data needs. The request from the NIA did not extend to prioritizing a research agenda, and the committee did not attempt to do so. The NIA and other research organizations dealing with questions about an aging society have many competing inputs and priorities, and since the committee did not evaluate them all, the committee felt it was inappropriate to be prescriptive about research priorities. Rather, these recommendations are designed to broadly inform NIA’s strategic research direction on the consequences of an aging society and to enable a more complete understanding of the relationship between population aging and the economy in the future.
DEMOGRAPHIC AND HEALTH MEASURES AND PROJECTIONS
1a. Improve methods of projecting mortality by age, sex, and socioeconomic characteristics. Projections of future life expectancy and mortality made by government agencies and demographers differ significantly, in part because projection methods and assumptions differ. The committee believes that increases in life expectancy will likely be more rapid than currently assumed in many projections. Research should consider whether (1) projections could be improved by explicitly taking into account trends in mortality related to smoking, obesity, and other behavioral factors, (2) projections could be improved by taking into account mortality differences by race/ethnicity and educational level (these are particularly important in light of current discussions about increasing the early and full retirement ages, which would have a larger proportional effect on expected years of retired life for groups with lower life expectancy and education), and (3) there is room for further improvement in projections based on formal demographic analysis of past trends and on trends within groups of countries.
1b. Investigate distributional aspects of the relationships between life expectancy, capacity to work, and income. Although the rising disparity in life expectancy across the income distribution has been documented, little is known about the causes of this widening. Similarly, little is known about trends in the capacity to work by income. Understanding the relationship between income distribution, capacity to work, and life expectancy—and having some basis for projecting these trends forward—is important to evaluating the distributional effects of proposals related to raising the eligibility ages for Social Security and Medicare.
1c. Better understand specific risk factors that are precedents of disability, including personal characteristics such as obesity and occupational hazards.
While disability rates among the older U.S. population have been constant over the past decade, disability rates among the working-age population have increased. Rising rates of disability with respect to mobility and related functions, combined with such secular changes as the dramatic increase in overweight and obesity among nonelderly and the tendency for underprivileged populations to drop out of school, suggest that future generations may fare less well than their predecessors. This may affect both the capacity of these future generations to participate in the workforce as they pass through middle age and beyond and also their need for personal care services as they get older.
1d. Quantify the effects of demographic change on state and local government budgets. The fiscal discussion in this report surveyed work to date on the effects of demographic change on state and local government budgets, but much remains to be done. In particular, future research should focus on the impact of demographic change on state and local expenditures for education and health services and on state and local tax revenues, as well as on accurate measurements of the future liabilities for pensions and health benefits for state and local workers (considerable work has been done on the pension side but not on retiree health benefits). Further, an analysis of the potential interactions between federal tax and entitlement policies and state budgets could be important to understanding the total impact of policy reforms: For example, would a delay in the full Social Security or Medicare retirement age have implications for state and local governments?
1e. Evaluate and extend measures and projections of disability/functional status described in this report and elsewhere. U.S. survey data suggest that declines in disability among older persons seen in the 1980s and 1990s have ceased during the past decade. These data generally focus on activities of daily living. Alternative conceptions of functional status sometimes suggest different results. Some researchers are questioning the use of chronological age as a basis for understanding health expectancy and people’s views of their own life prospects and have developed alternative measures that address concepts of old-age dependency and work potential. The committee believes the time is ripe for a broad evaluation of different approaches, with an eye to building on work described in this report.
WORK CAPACITY AND INCENTIVES FOR
A LONGER WORKING LIFE
2a. Examine what past relationships between health status, age, and economic incentives for continued work suggest about the likely path of labor
force participation rates for older (60–75) workers over the next 30 years. Labor force participation rates have already reversed their long-term decline for men. The average retirement age has risen for women as well, and it appears that the older population is not becoming any less healthy. In view of these and other trends, can we better project labor market activity? There are several elements to this question: (1) What fraction of the older population will have health limitations that prevent them from earning much income at age 62, 65, 68, or even 70? (2) How responsive is labor supply to financial incentives to work? (3) Will the labor market deliver opportunities for older workers that will enable the labor force participation rate to rise significantly? and (4) How might changing labor force patterns among women contribute to the overall labor force participation rate? Part of the research challenge in this area is to distill and synthesize the large quantity of empirical evidence on labor supply response to various factors.
2b. Assess the development and adoption of technologies that enhance the capacity of individuals with specific disabilities to participate in the workforce. There are a number of ways in which policy can affect health and disability in the population. One way in particular would be to encourage the development of technologies, including person-based biomedical advances as well as workplace-based devices, that can help improve the health and capacity of individuals with various types of disabilities to perform job-related functions.
2c. Undertake specific analyses of how work at older ages has been and could be facilitated, including demand-side factors. The analyses within this report strongly suggest that more people will be working to older ages than has been the case in past decades. How long people work will depend not only on health trends and pension incentives, but also on the demand for older workers, opportunities for training, retraining, and continuing education, and the flexibility of work at older ages. As our society moves toward later retirement ages, it will be important to (1) assemble disparate information on the types of arrangements that older workers would prefer; (2) understand the arrangements that firms have created or are willing to think about; and (3) evaluate the effect of any trials.
2d. Study the effects of the health reforms enacted in 2010 (known collectively as the Affordable Care Act, or ACA) on labor force participation. The ACA could have important effects on the retirement decision of older workers and on labor force participation decisions more generally. On the one hand, the ACA will make it easier and less expensive for workers who retire early to purchase health insurance on their own and might thereby
encourage more early retirement. On the other hand, it is possible that some potential workers currently remain out of the labor force in order to be eligible for Medicaid or Medicare (through the Social Security disability program). The increased affordability and availability of health insurance under the ACA could raise the labor force participation of these workers. Understanding the magnitude of each of these effects will contribute importantly to our ability to gauge the likely macroeconomic effects of demographic change and to evaluate the likely impacts of various policy reforms.
CHANGES IN CONSUMPTION AND SAVING
3a. Rethink how consumption “needs” change at retirement and how they evolve over the course of the retirement period. One of the most difficult challenges in evaluating the findings on retirement resource adequacy is deciding what the benchmark should be for postretirement consumption. Is an 80 percent replacement rate of income sufficient? Is there a substantial opportunity to substitute home production for market purchases, and do job-related costs represent a substantial share of preretirement consumption spending? How do these factors vary with age and duration of retirements that may last 30 years or more? There have been some studies directed at these issues, but further refinement would contribute to a better understanding of saving adequacy for future cohorts of retirees.
3b. Elucidate the interactions between private saving and government policy. The macroeconomic effects of policy changes depend on how individuals respond to them. In particular, How would households respond to reductions in Social Security or Medicare benefits? A wide empirical literature exists on how the retirement decision is affected by the eligibility age for public pensions, but much less is known about saving behavior and retirement age, or saving behavior and Medicare generosity. Of particular interest, given the likelihood that it will occur, would be the effect on private saving decisions of increased means-testing of benefits.
3c. Evaluate how sensitive retirement preparedness results are to scenarios for financial market returns and changes in health care cost growth. If house prices remain flat in real terms for 10 years, for instance, how would this affect retirement readiness results in 2025? What about a low real return on equities for a decade—a repeat of the 2000–2012 experience, for example? This should be a relatively straightforward set of calculations using extant data from national surveys (e.g., the Health and Retirement Study and the Survey of Consumer Finances) in tandem with projections of asset market movements.
3d. Explore whether and for whom personal saving and labor market attachment will increase with longer lifetimes. If people work longer as their lives are extended, society might be able to pay for old-age income and health care programs targeted at those who cannot self-finance. If people also save more privately, this could reduce their need to draw on public safety-net public programs in old age. But the extent to which various subgroups in the population can and will save more and work longer, as the life cycle lengthens, requires additional research.
MODELING AND DATA
4a. Promote modeling efforts to simulate how demographic aging may interact with changing patterns of health status, labor force participation, saving behavior, and capital market movements. New models of the macroeconomic effects of population aging would allow us to better characterize the sensitivities of projections and the interactions between macroeconomic variables. Resulting simulations would help identify the most important policy levers available now and in the future to influence the adequacy of retirement income. Such analyses can also suggest where coordinated policy actions across several domains can be most productive (e.g., improvements in health and functional status can impact labor force participation and productivity) and reduce the costs of adjustment to demographic aging.
4b. Improve existing models of life-cycle saving and decumulation, and test them with data that link individuals with administrative information on saving, investments, and retirement drawdowns. Longitudinal data collection efforts are critically important for tracking and modeling how older households prepare for, and move into, retirement. Better use of the full range of longitudinal data could provide a real-time means to study policies that influence retirement security. Research also is needed on (1) understanding how improving financial literacy can improve retirement security, including how economic and noneconomic (psychological and sociological) factors interact to affect retirement security, and (2) the effectiveness of financial products and related innovations (e.g., longevity risk pooling, inflation and survivor bonds, and long-term care insurance) that have been and might be developed to help people better manage key risks.
4c. Support and broaden ongoing multidisciplinary survey research. The committee strongly endorses the need for continuing longitudinal surveys that include in a single comprehensive instrument questions on health, functional status, retirement behavior, income, pensions, savings, well-being, planning, and related matters. The Health and Retirement Study (HRS)
does exactly this and has been enormously valuable for research on the economics of aging. The availability of comparable longitudinal surveys in many developed and developing nations has greatly enhanced the value of the HRS.
4d. Improve the development and use of macro data. While data at the individual and household level are extremely valuable, it also is useful to have macro-level data on many aspects of economic behavior, disaggregated at least by age and preferably by other variables as well. Data of this sort often can be constructed from existing administrative and survey sources such that no new data collection efforts are needed.