Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Introduction Many government policies and programs that affect the economic well-being of Americans are under scrutiny to assess their costs and benefits. In 1993-1994, the key public debate was on how to provide health care at a reasonable cost; in 1995-1996, the debate focused on the costs and benefits of federal and state programs to aid the poor. Although one cannot forecast with certainty what the next key public issues will be, it seems very likely that one of the major debates will be about the income security of current and future generations of retirees and the costs to workers and the economy to attain that security. Significant demo- graphic changes the aging of the baby boom generation and the relatively small size of the subsequent baby bust generation coupled with fiscal pressures on the federal government are two factors that are raising concerns about the prospects for the U.S. Social Security system. Just as important are concerns about em- ployer pensions and other benefits, personal savings rates, and the adequacy of retirement income in the face of rising health care costs. As policy makers wrestle with the issues, they will need estimates or projec- tions of the likely costs and other effects of continuing current retirement- income-related policies in comparison with one or more alternative policies. Anticipating these needs, the Pension and Welfare Benefits Administration (PWBA) in the U.S. Department of Labor asked the Committee on National Statistics at the National Research Council to establish a Panel on Retirement Income Modeling. The National Institute on Aging, Pension Benefit Guaranty Corporation, Social Security Administration, and TIAA-CREF also provided sup- port for the project. The charge to the panel was to recommend improvements to 10
INTRODUCTION 11 retirement-income-related research, data, and models that analysts can use to help answer decision makers' questions and inform the policy debate. Without reasonably good estimates from models of the short-run and long- run implications of current policies and alternatives to them, decision makers run the risk that the proposals they adopt may be ineffective or, worse, counterpro- ductive. Yet the task of developing reliable estimates is far from easy, and it becomes even more challenging as any proposed set of policy changes becomes more complex and far reaching. The experience of health care reform underscores the problems for decision making when estimates of the effects of alternative policy proposals are highly uncertain. During the course of the 1993-1994 debate, different projection mod- els produced widely differing estimates of the effects of the same health care reform plan, and those estimates were often hard to reconcile. The lack of agreed-upon estimates was not the reason for the failure of Congress to enact any of the reform proposals, but it undoubtedly made the deliberative process more difficult (see Mann and Ornstein, 1995:Chs. 4,7,8~. Much of the uncertainty in the estimates of alternative health care reform proposals was because there was simply no experience with key provisions of the Clinton administration and other plans. Analysts could only guess at the likely effects of such changes as establishing statewide health insurance purchasing cooperatives or requiring employers to provide health insurance for their work- ers. Yet investments in relevant data and associated research and modeling efforts prior to 1993 could likely have reduced some of the uncertainty and, more generally, provided an improved capability for answering decision makers' ques- tions. In particular, it appears that the lack of needed data and behavioral re- search handicapped the effort to develop useful models for health care policy projections (see Bilheimer and Reischauer, 1996~. In seeking advice about retirement-income-related data, research, and mod- els, PWBA sought to learn from the health care reform experience. PWBA asked our panel to identify investments in research, data, and models that would better prepare analysts to respond to decision makers' likely information needs about retirement income. RETIREMENT INCOME SECURITY To adequately assess the economic well-being of older Americans, it is important to consider all sources of income and wealth and the risks facing each source, as well as the overall risks to income security posed by increasing health care costs. It is also important to assess the associated costs to workers, their families, and the economy of policies that are designed to further retirees' well-being. A broader interpretation of retirement income security could include such issues as whether there are adequate labor market opportunities for retirement-age people
2 ASSESSING POLICIES FOR RETIREMENT INCOME who want to work on a full-time or part-time basis and whether there are adequate training and rehabilitation facilities for disabled older people. Sources of retirement income include Social Security, employer-provided pensions, other transfer programs (e.g., public assistance, public and private dis- ability payments, and workers' compensation), personal savings, inheritances and other transfers among family members, and post-retirement earnings (e.g., from part-time work). A major potential drain on retirement income is expendi- tures necessary to maintain or restore health; these costs, if they continue to grow as in the past, may significantly undercut the adequacy of income for other needs. Public policies in many areas affect one or more components of retirement income security, both directly and indirectly through their effects on employer and employee behavior and on the general level of economic activity. These policy areas include: · the Social Security system both provisions of the payroll tax and the benefit structure, including retiree, survivor, and disability benefits; · the federal personal and corporate income tax code provisions that af- fect employer-provided pensions and health care benefits, such other forms of saving as housing and Individual Retirement Accounts (IRAs), and bequests; · the Medicare and Medicaid systems and public assistance programs, such as Supplemental Security Income (SSI) for the poor elderly and disabled; · government regulations that affect employer work force practices, such as laws prohibiting age discrimination in hiring and mandatory retirement and the Americans with Disabilities Act, and those that affect pension benefits, such as the Employee Retirement Income Security Act (ERISA); · government agency (and employer) efforts to educate workers about their pension benefits and investment options; and · overall macroeconomic policies that affect inflation, employment, and real wage growth, which, in turn, have major implications for retirement income security. Many separate agencies are responsible for each of these policy areas: the U.S. Department of the Treasury; the Health Care Financing Administration in the U.S. Department of Health and Human Services (HHS); the Pension and Welfare Benefits Administration in the U.S. Department of Labor (DOL); the Pension Benefit Guaranty Corporation; and the Social Security Administration. In addition, other agencies are involved in retirement-income-related data collec- tion and research: the Agency for Health Care Policy and Research, National Center for Health Statistics, and National Institute on Aging in HHS; the Bureau of Labor Statistics in DOL; the Bureau of the Census in the U.S. Department of Commerce; and the National Science Foundation. The debate over retirement income security in the next few years will likely consider changes to many of the policies listed above. Areas of particular interest
INTRODUCTION 13 will be employer pension regulation, tax incentives for personal savings, and Social Security; health care policy will also be of continuing concern. Decision makers will likely want estimates for a wide range of questions about proposed changes in these areas, as illustrated below. Employer Pension Regulation There has been growing interest in pension simplification. Recent legislation simplified pension regulations for small busi- nesses to encourage them to offer certain types of pension plans. In 1993, almost two-fifths of private full-time workers were employed by businesses with fewer than 100 full-time employees, and fewer than one-fourth of these workers partici- pated in pension plans (Pension and Welfare Benefits Administration, 1994:Table Bib. To evaluate the effects on retirement income security of the new legisla- tion and of future laws or rules to simplify pensions one would like estimates of the number and proportion of small businesses that will respond by setting up particular types of plans; the number of workers in those businesses who will participate; and the amount of contributions, in dollar terms and as a proportion of earnings, that employers and workers will make. One would also like answers to such questions as: Will low-wage workers participate to the same extent as higher wage workers? Single workers as much as workers with dependents? Small businesses in certain industries as much as small businesses in other indus- tries? Will there be any short- or long-term effects on total employment in small businesses? What will be the effects on workers' pension coverage during their working careers and on the income they receive in retirement? Tax Incentives for Personal Savings Proposals are often floated to induce more people to set up and contribute to IRAs. Recent legislation increased the limit on IRA contributions for spouses who do not earn income. Other proposals are to allow higher income people to defer taxes on contributions to IRAs and to allow IRA holders to use their funds for such purposes as medical care or college tuition, in addition to retirement. Important questions about such proposals in- clude: Will they encourage an increase in total personal savings, or will people simply put money into their IRAs that they would have saved in other forms? Will people who are not now saving set up IRAs, or will most new accounts be set up by people who are already saving significant amounts in other ways? Will the amounts put into IRAs be available for retirement or will people spend them earlier? What will be the revenue losses (or gains) to the government of tax deferrals on IRA contributions or earnings? Social Security Reform Included in some proposals to resolve the Social Security financing problem for the long term is "privatization" of part or all of the old-age, survivors, and disability insurance contributions (some members of the Social Security Advisory Council recommended partial privatization; see Quinn and Mitchell, 1996~. In this approach, all or a fraction of the current payroll tax
4 ASSESSING POLICIES FOR RETIREMENT INCOME contribution would be put into an individual retirement account for each worker. Key questions about privatization include: What are the implications for equity of Social Security taxes and benefits across generations? How can or will the social insurance components of the current system be funded (the provisions to support disabled workers and dependents of deceased workers and to provide disproportionately more benefits to workers with low contributions due to low earnings or interrupted work careers)? How will individual accounts be adminis- tered? What will be the administrative costs of such a system in comparison with the current system? What regulations will be needed to protect workers' invest- ments (particularly for proposals that allow worker choice)? How will the avail- ability of indexed Treasury bonds as an investment option affect the return on and security of workers' accounts? What restrictions, if any, will be placed on the use of workers' accounts when reaching retirement age, such as requiring single or joint life annuitization, possibly indexed for inflation? What will be the effects of those restrictions? How can a transition be made between the current and the new systems? What effects will there be on employer pension plan coverage and benefits, which, in many instances, are linked to Social Security provisions? What effects will there be on capital markets, and what will be the consequences for workers' retirement income security, overall and for workers with different characteristics? THE REPORT The above examples illustrate the breadth and depth of questions that need to be addressed if choices among retirement-income-related policy options are to be made on an informed basis. The challenges of developing reliable estimates to respond to these and other questions about the likely effects of alternative poli- cies are major, and not all, or even many, of the questions may be answerable with currently available or foreseeable data and analysis. The purpose of our report is to identify improvements in data, research, and modeling that can better enable analysts to respond to decision makers' questions and information needs and that can reduce the uncertainty that is inevitable in projections. Chapter 2 provides a framework for retirement income modeling. It briefly reviews concerns that have prompted increasing attention to policy issues in this area, policy options that are likely to be considered, outcomes on which to evalu- ate various options, and challenges to developing reliable projections of out comes. Chapter 3 summarizes the state of basic research knowledge in key areas, including employers' decisions about pension and health care benefits and their demand for older workers, people's choices about savings, consumption, and retirement, and factors that affect the size and composition of the older popula- tion and the health care needs and costs of older people. Drawing heavily on a set of commissioned papers in Hanushek and Maritato (1996; see Appendix A), the
INTRODUCTION 15 chapter identifies key questions that have not been satisfactorily addressed and lists priority topics for research and behavioral modeling. Chapter 4 recommends improvements in data sources that are essential to understand key behaviors and to develop good projection models to answer policy makers' questions. Indeed, we argue that, in light of limited resources, agencies must give priority to investments in data and analytical research. Chapter 5 offers a vision for improved policy projection capabilities when better data and research knowledge become available. Finally, Chapter 6 recommends ways to facilitate interagency coordination and the involvement of the private sector and academia in retirement-income- related data collection, research, and modeling. Appendices provide information on relevant data sets and projection models and on some of the problems with available models. An important distinction that we make throughout the report is between analytical or behavioral models (which are the focus of Chapter 3) and projec- tion models (which are the focus of Chapter 5~. Very roughly, analytical models are research tools, and projection models are policy tools. Analytical models are intended to answer "why" and to determine the strength of key relationships. As an example, we include in the analytical category models that are used to estimate the probabilities of specific behavioral responses (e.g., increased savings, de- creased work effort, increased employer contributions to benefit plans) as a func- tion of characteristics of the individual or the employer, policy features, and other factors (e.g., interest rates). Projection models are intended to answer "what if" questions and to project future outcomes. We include in the projection category models that are used to estimate the likely costs and other effects over time of current and alternative policies. Projection models may use a variety of informa- tion, including estimated behavioral probabilities from analytical models. Both kinds of models are important to develop for retirement-income-related policy analysis purposes. Analytical and projection models vary in scope, complexity, and modeling strategy, and, consequently, in their costs and benefits. For example, micro- simulation is an approach to projection modeling that operates at the individual decision unit level. Microsimulation models can provide detailed information about the likely effects of policies for particular kinds of people or employers, but they have heavy data requirements and can be difficult to implement and under- stand. Other kinds of projection models, such as time-series models, cell-based models (e.g., population projection models), macroeconomic models, and com- putable general equilibrium models, have their own advantages and drawbacks. The same is true of different types of analytical models, such as reduced-form models and static and dynamic structural models. To the extent possible, we endeavor to consider cost-benefit tradeoffs in our recommendations for models, data, and research. Decision makers need good
16 ASSESSING POLICIES FOR RETIREMENT INCOME policy analysis and projection capabilities, but resource constraints dictate that priorities be set and choices be made. Finally, we consider decision makers to be not only officials of the federal government who are concerned with pension regulation, taxation, Social Secu- rity, and other relevant programs, but also federal, state, and local officials and private-sector executives in their capacity as employers. Pensions and other benefits and, more generally, hiring and compensation practices, are important factors in both retirement income security and employer costs. Improved data, research knowledge, and analytical and projection models can help employers, as well as policy makers, make more informed decisions.
