Employment Support for the Transition to Retirement: Can a New Program Help Older Workers Continue to Work and Protect Those Who Cannot? David C. Stapleton Center for Studying Disability Policy Mathematica Policy Research, Inc. Employment Support for the Transition to Retirement: Can a New Program Help Older Workers Continue to Work and Protect Those Who Cannot? David C. Stapleton Center for Studying Disability Policy Mathematica Policy Research, Inc. AARP’s Public Policy Institute informs and stimulates public debate on the issues we face as we age. Through research, analysis and dialogue with the nation’s leading experts, PPI promotes development of sound, creative policies to address our common need for economic security, health care, and quality of life. The views expressed herein are for information, debate, and discussion, and do not necessarily represent official policies of AARP. #2009-05 April 2009 © 2009, AARP Reprinting with permission only AARP Public Policy Institute 601 E Street, NW, Washington, DC 20049 http://www.aarp.org/ppi ACKNOWLEDGMENTS This paper has greatly benefited from comments on earlier drafts by Richard Burkhauser, John Gist, Howard Iams, Gina Livermore, Janet McCubbin, Ellen O’Brien, Virginia Reno, Sara Rix, Jody Schimmel, Eugene Steuerle, and Craig Thornton. I am grateful for the financial support of AARP’s Public Policy Institute. The contents of the paper do not represent the official position of AARP. I retain sole responsibility for the opinions expressed and any errors. ii TABLE OF CONTENTS ACKNOWLEDGMENTS ................................................................................................... ii ABSTRACT ...................................................................................................................... v EXECUTIVE SUMMARY ................................................................................................ vii A New Approach: Employment Support for the Transition to Retirement ................ vii Who Would Be Eligible for ESTR?........................................................................... viii How Many Would Be Eligible?................................................................................... ix How Would the Eligibility Determination Process Work?.......................................... ix What Benefits Would ESTR Provide?.......................................................................... x How Long Would Benefits Last? ................................................................................ xi What Would ESTR Cost? ............................................................................................ xi How Would the ESTR Program Be Administered and Financed? .............................. xi Conclusion ................................................................................................................... xi THE POLICY PROBLEM.................................................................................................. 1 ELIGIBILITY ..................................................................................................................... 7 Target Population Concept ........................................................................................... 7 Operational Definition of the Target Population .......................................................... 7 Size of the Target Population and Number of Participants......................................... 10 Eligibility Determination ............................................................................................ 11 BENEFIT OPTIONS........................................................................................................ 11 Disability Screening and Expedited SSDI Determinations ........................................ 12 Benefit, Employment, and Financial Counseling ....................................................... 13 Wage Subsidy ............................................................................................................. 14 Subsidies/Allowances for Extraordinary Work-Related Expenses............................. 15 Health Insurance Subsidies ......................................................................................... 16 Health Risk Reduction Services.................................................................................. 18 Extended Unemployment Benefits ............................................................................. 19 Supplemental Security Income and Medicaid ............................................................ 20 iii COST .............................................................................................................................. 21 FINANCING AND ADMINISTRATION ........................................................................... 23 CONCLUSION ................................................................................................................ 24 APPENDIX: POLICY BACKGROUND........................................................................... 28 Retirement Policy Reform .......................................................................................... 28 Disability Policy Reform ............................................................................................ 31 Trade Adjustment Assistance ..................................................................................... 34 REFERENCES................................................................................................................ 37 iv Employment Support for the Transition to Retirement ABSTRACT There are compelling reasons for policy reforms that would encourage later retirement, but one important drawback to many reforms is that they would adversely affect the household incomes of the many workers who experience large involuntary earnings losses as they approach retirement. Employment Support for the Transition to Retirement (ESTR) is a set of supports designed to encourage and help such workers increase their earnings and postpone reliance on their retirement benefits until the benefits are larger. This paper describes the need for such a program; presents options with respect to eligibility and benefit design; considers the potential cost, financing, and administration of the program; and compares the approach embodied in ESTR with other approaches designed to address the needs of those most adversely affected by retirement policy reforms. v This page intentionally left blank. Employment Support for the Transition to Retirement EXECUTIVE SUMMARY Many working Americans expect to work until they reach the Full Retirement Age (FRA) for Social Security (currently 66), and many expect to continue working into their “retirement years.” Many, however, are forced to retire early due to poor health, long-term unemployment, or other events. To address the economic security of older workers whose working lives are cut short, Social Security allows workers to claim reduced benefits beginning at age 62, the Earliest Eligibility Age (EEA). A worker or retiree needs to be fully insured to receive these Social Security benefits, but no other restrictions apply. This early retirement option provides a solid floor of protection for workers who are forced by circumstances to retire early, but it also provides benefits to people who may not have a pressing “need” for them and who might well be better off waiting. Policies to extend working lives—and delay the claiming of Social Security benefits—are viewed as a necessity by many experts who have grave and growing concerns about workers’ retirement security and Social Security’s finances. One of the leading proposals to extend working lives is to increase the earliest eligibility age in Social Security. Analysts contend that the surest way to keep older workers in the labor market is to stop “sending the signal” that retirement at age 62 is an option. The major problem with raising the EEA is that doing so would inflict very real hardship on some older workers. Understandably, policy makers might be reluctant to raise the earliest eligibility age unless a policy to attenuate hardship is in place. In a research paper for the AARP Public Policy Institute, David Stapleton of Mathematica Policy Research, Inc. proposes a new program could help “break the deadlock” that stymies efforts to adopt policies that encourage later retirement. Stapleton describes a program—which he calls Employment Support for the Transition to Retirement (ESTR)—that could address the harm that an increase in the earliest eligibility age might inflict on some older workers, especially those who are seriously ill or disabled, who experience long-term unemployment, or who remain employed but who experience wage or income loss for reasons beyond their control. Stapleton identifies a number of circumstances that he says would call for expanded assistance for workers nearing retirement if early benefits were no longer available, and calls for a wide range of benefits, tailored to individual need—including wage subsidies and other work supports, health insurance subsidies, disability benefits, extended unemployment benefits, and employment counseling. The elements of ESTR are not new; they can be found in existing federal and state programs. What is new is the idea of a substantial and coordinated expansion of these elements in the context of retirement policy reform. A NEW APPROACH: EMPLOYMENT SUPPORT FOR THE TRANSITION TO RETIREMENT In addition to Social Security’s early retirement benefits, an existing safety net provides assistance to unemployed older workers, to workers displaced by globalization, and to disabled workers through the Unemployment Insurance program, through the Trade Adjustment Assistance (TAA) and Alternative Trade Adjustment Assistance (ATAA) programs, and through the Social Security Disability Insurance (SSDI) program and the Supplemental Security Income (SSI) program. One option would be to expand these vii Employment Support for the Transition to Retirement programs to meet the needs of older workers if early benefits were reduced. Stapleton proposes to integrate these programs in his new approach. Stapleton envisions a program that would provide assistance to workers who experience large involuntary earnings losses as they approach age 62. The idea behind the program— Employment Support for the Transition to Retirement—is to provide targeted assistance to people likely to claim early retirement benefits because they have no other option. ESTR would also address limitations in the Social Security Disability Insurance program. Stapleton says ESTR would expedite the disability determination process for older workers with disabilities and provide partial disability benefits to people who can work longer. Perhaps the most fundamental question about ESTR is the difficulty of drawing a line between voluntary and involuntary reductions in the worker’s earnings. Nevertheless, Stapleton suggests that a targeted and individualized program perhaps can be designed that may be cost effective and not too administratively burdensome. Stapleton identifies key program design choices and briefly describes some of the consequences of different choices, but he leaves many questions open to further debate in this first attempt to think about what a program might look like. WHO WOULD BE ELIGIBLE FOR ESTR? Under the proposed approach, a comprehensive set of wage and employment supports would be available to certain older workers, beginning at, say, age 55. Older workers would qualify for support if they were fully insured for Social Security benefits and if they experienced adverse circumstances such as: (1) a long-lasting medical condition or disability that interferes with work, reduces earnings or results in substantial financial expenses; (2) a change in the local labor market that substantially reduces the worker’s earnings and earnings prospects; or, (3) a substantial increase in the care needs of a dependent or immediate family member with a significant medical condition that, for example, imposes time demands on the worker that are incompatible with continued full- time employment at his or her current job. To qualify, the worker would have to demonstrate that his or her annual earnings are below a threshold value—defined as a percentage of some measure of the worker’s past earnings—having declined for reasons that are beyond his or her control. The size of the percentage reduction in earnings and the earnings base, used to measure the percentage reduction, will be key elements in determining how many workers become eligible for ESTR. One option for the earnings base is to use Social Security’s Average Indexed Monthly Earnings (AIME) value as of a certain age (55, for example). AIME is the measure of past earnings that is used to determine the level of benefits under the current program. The eligibility ages for ESTR are also a key design choice, with important implications for both program effectiveness and cost. On the one hand, Stapleton explains, eligibility could be limited to workers age 62 and older, on the grounds that retirement policy reforms are not likely to have an adverse impact on the target population until they reach age 62. On the other hand, eligibility at an earlier age, such as 55, would make ESTR available to many workers in the target population when they first experience earnings loss, at the point when support is most likely to have a positive impact on earnings and reduce SSDI entry. viii Employment Support for the Transition to Retirement All existing SSDI beneficiaries meeting the ESTR minimum age requirement would be allowed to exit SSDI and enroll in ESTR. Former SSDI beneficiaries whose benefits were terminated because of earnings (and who have not experienced medical improvement) would also be eligible. This option is intended to entice SSDI recipients back into the labor force and help them increase eventual retirement benefits. HOW MANY WOULD BE ELIGIBLE? A few hundred thousand workers in each birth-year cohort experience an involuntary earnings reduction and substantial economic hardship as they approach 62 (without having entered SSDI). The number of these workers who become eligible for ESTR will depend on the stringency of the eligibility criteria (that is, what constitutes an involuntary earnings reduction, and how large it must be) and how those criteria are implemented. Stapleton estimates reasonable eligibility criteria could lead to 150,000 to 300,000 workers between the ages of 55 and 62 becoming eligible for ESTR in a typical year. If the EEA were increased, additional workers would become eligible between age 62 and the new EEA. The number of eligibles who enroll will presumably be smaller, and will depend on the value of the benefits, time limits and administrative choices. If participation is high (e.g., 80 percent), there could be 125,000 to 250,000 program entrants each year. By comparison, 715,000 62-year-olds were receiving Social Security retirement benefits in June 2008, and another 326,000 were enrolled in SSDI. If the minimum eligibility age is 55 and entrants at any age can remain in the program until they reach the EEA there could be as many as 750,000 to 1.6 million participants at any time. If, instead, a two-year limit on benefits is imposed, the number of participants each year will be approximately twice the number of entrants—perhaps 250,000 to 500,000. Analysis of likely per enrollee costs suggests that the number of participants in a typical year will have to be at the lower end of the figures in the previous paragraph for the program to be fiscally attractive from the government budget perspective. HOW WOULD THE ELIGIBILITY DETERMINATION PROCESS WORK? Stapleton envisions a web-based application process, with eligibility determination based on administrative records, primarily Social Security date-of-birth and earnings records. Given the complexity of the information requirements for determining adverse circumstances, the eligibility-determination process would require substantial application assistance. A web-based application program would be developed to help applicants and their advisors gain access to their Social Security data, determine what additional information is required, assemble all needed information, and submit the application electronically. Such a system would have to be supplemented by paid counselors available to those not capable of navigating the application process alone. The application process could be substantially automated. Intake workers would address problems with administrative data and conduct simple consistency and validity checks on the information submitted by applicants (to the extent that such checks cannot be fully automated), and could return claims for lack of adequate documentation, but might not otherwise play an adjudicative role. Establishment of the National Health Information Network would facilitate the process. To minimize administrative burden and maintain short application processing times, program stewardship could potentially rely heavily on ix Employment Support for the Transition to Retirement a system of random audits and penalties rather than on comprehensive validation of evidence. Applicants could be required to submit, document, and attest to the veracity of this information, and documentation could include forms to be signed by medical professionals, employers, and perhaps others. WHAT BENEFITS WOULD ESTR PROVIDE? Stapleton envisions a targeted program with a rich set of benefits tailored to individual circumstances. The benefits a person receives would depend on such factors as their employment and earnings history, their health care needs and access to insurance, and their disposable income and available resources. Eight core benefits are described; they include:  Disability screening and expedited SSDI determinations. Entrants are screened for the likelihood of SSDI eligibility. Workers with severe disabilities that prevent work would be enrolled in SSDI. Those with disabilities who are able to work would be enrolled in ESTR, with expedited entry into SSDI if work efforts fail.  Counseling. Benefits counseling, employment counseling and financial counseling to help older workers develop an economic security plan for the remainder of their lives.  Wage subsidy. Those working more than a minimum number of hours per week would be entitled to a time-limited wage subsidy equal to a fraction of the gap between current earnings and an earnings target up to a maximum, based on the participant’s past earnings.  Subsidies/allowances for work-related expenses. These would address extraordinary work-related expenses that the worker or worker’s employer must incur, including expenses for the care of a dependent with significant disabilities, job- search costs, and relocation expenses associated with a new job.  Health insurance subsidy. A subsidy could be made available to participants as well as to employers of participants who offer employee coverage. The subsidy could be in the form of a new Medicare buy-in, an expansion of the Medicaid Buy-in, a refundable tax credit for private health insurance, or publicly funded stop-loss coverage for private plans.  Health risk reduction services. Designed to improve health and longevity, reduce the demand for high cost health care services, increase productivity, and make it easier to work.  Extended unemployment benefits. Unemployed participants could be entitled to additional weeks of unemployment benefits, possibly including incentives to return to work.  Supplemental Security Income (SSI) and Medicaid. Those not able to achieve sufficient economic security with other ESTR benefits would be eligible for SSI and Medicaid under current rules for elderly SSI and Medicaid recipients. x Employment Support for the Transition to Retirement HOW LONG WOULD BENEFITS LAST? A 24-month time limit on eligibility for ESTR could limit the potentially high cost of offering eligibility at age 55. Any time limit could also be applied selectively to various ESTR benefits. WHAT WOULD ESTR COST? The cost of ESTR will depend on how many people are eligible, how many enroll and what benefits they receive. If the Social Security early eligibility age is increased from 62 to 64 and ESTR is available on a time-limited basis to those age 55 and over who meet the other ESTR eligibility criteria, Stapleton suggests that the number of participants could be in the range of 250,000 to 500,000. Program cost per enrollee will depend on specific features of ESTR, but “to be politically viable,” it would have to be below the current average annual federal expenditure per SSDI beneficiary, which is $28,000 (the average cost of Social Security benefit payments and Medicare costs). If the annual cost per enrollee is $15,000, annual benefit costs would range from $3.75 billion to $7.5 billion per year. Administrative costs are likely to be large relative to program costs—perhaps 20 percent of program costs, in which case total cost would range from $4.5 billion to $9 billion. Equally important to making the case for reform is the impact of retirement policy reform, incorporating ESTR, on payroll and federal income tax revenue. Stapleton suggests that an increase in the earliest eligibility age or other retirement policy reforms could generate enough federal revenue to pay for a program at the lower end of this range, but perhaps not at the higher end. HOW WOULD THE ESTR PROGRAM BE ADMINISTERED AND FINANCED? Currently, several federal and state agencies are responsible for administering various components of ESTR benefits for other populations, and financing comes from a variety of revenue streams. ESTR could potentially be implemented through a coordinated set of changes to these agencies’ responsibilities and funding streams at both the federal and state level. A more desirable model would establish a single point of contact for workers, in local offices run by a state agency under the oversight of a federal agency. CONCLUSION Arriving at the specifics of a policy to encourage and support longer working lives for older workers will require much additional examination and debate. Stapleton observes, however, that increases in longevity—and the expectation of future increases—make retirement policy reforms seem inevitable. Many European countries have an automatic link between public retirement benefits and expected longevity, although these countries provide greater supports for older workers who experience job loss. Strengthening supports for older workers in this country, Stapleton thinks, might be key to moving forward with retirement policy reforms. xi Employment Support for the Transition to Retirement ESTR’s approach is consistent with current thinking about programmatic reforms to address the economic security needs of all working-age people with disabilities. ESTR is also broadly consistent with a trend toward public policies that provide assistance to disadvantaged workers through temporary subsidies and other supports that increase the incentive and ability to work. xii This page intentionally left blank. Employment Support for the Transition to Retirement THE POLICY PROBLEM There are compelling reasons for the federal government to adopt policy reforms that would encourage workers to stay in the labor force longer than they do under current policy. An increase in the Earliest Eligibility Age (EEA) for Social Security Retirement (SSR) benefits is perhaps the most prominent example of such a policy change, but there are many others.1 The driving force behind proposals for such reforms is the increase in average longevity after age 62 (the current EEA) coupled with the growing number of individuals who remain healthy and vigorous beyond that age. Increased longevity means that today’s average new SSR entrant will receive benefits for many more years than individuals who entered SSR when the program was first established. Further, if retirement patterns do not change, the size of the population consuming goods and services relative to the size of the population producing goods and services will continue to grow. Similarly, the population of those receiving SSR and Medicare benefits is growing relative to the size of the population generating revenues for those programs via payroll taxes on their earnings. These changes are often cited as the primary cause of the projected insolvency of the Social Security Trust Fund and as a major contributor to the projected insolvency of the Medicare Trust Fund. The beginning of the retirement of the Baby Boom generation in January 2008 portends near-term acceleration of these trends. One major impediment to retirement policy reforms is the potential harm that many proposed reforms might inflict on workers for whom delayed retirement would be a significant hardship, most notably those with work-limiting health conditions as they approach the current EEA. Mounting evidence that gains in longevity are not distributed uniformly suggests that there are many older workers who are not substantially benefiting from advances in health. Most recently, Waldron (2007) has documented that longevity gains for men since the early 1900s have been concentrated among those in the top half of the earnings distribution during their prime work years.2 Several earlier studies cited by Waldron have documented large and widening longevity differences for men across levels of education. National statistics also show significant differences in life expectancy at age 65 by race.3 The findings of research on increased longevity suggest that gains in health among older workers have also been concentrated among those with relatively high earnings and education, but no research has yet tested this hypothesis. There is, however, substantial evidence that the number of workers experiencing health-related work limitations as they approach retirement continues to be large, especially among those with relatively low levels of education. Substantial numbers of other workers approaching retirement experience involuntary earnings losses for other reasons, most notably long-term or permanent layoffs, and at 1 A summary of other proposed retirement policy reforms appears in the appendix. 2 See the appendix for further discussion of the evidence on disparities in longevity gains. 3 Life expectancy for white males at age 65 is 17.2 years compared to 15.2 years for African American males (National Center for Health Statistics 2007). 1 Employment Support for the Transition to Retirement least some such workers might wind up considerably worse off under policy reforms that delay their access to retirement benefits. The purpose of this paper is to contribute to the discussion on retirement policy reform by presenting the framework for a program that would address the harm that many reforms might inflict on workers who experience involuntary earnings loss as they approach the current EEA. Development of such a program could help break the deadlock that has stymied efforts to adopt policies that encourage later retirement. Several other countries have moved forward with policies to encourage later retirement, and their reforms have often included components to address the needs of older workers who experience involuntary wage losses (Martin, 2008). The latter type of reform has received inadequate attention in the policy debate in the U.S. In the remainder of this section, we review the current federal and federal-state programs that provide economic security for workers who experience involuntary wage loss, with special attention to older workers. We then review the evidence on the extent to which older workers experience involuntary wage loss, and its consequences for household incomes. In subsequent sections, we sketch and discuss a policy option designed to address the needs of such workers as a component of broader retirement policy reforms. Under current law, all workers insured for Social Security can increase their household incomes at age 62 by claiming SSR benefits. Monthly benefits will be permanently lower than they would be if the worker postponed claiming until attainment of the Full Retirement Age (FRA – currently 66), by an actuarially fair amount. For workers who have experienced involuntary earnings loss, whatever the reason, early claiming might be the only option available to minimize a decline in household income. The value of this option to such workers is already being eroded by a retirement policy reform that is currently being implemented: a gradual increase in the Full Retirement Age (FRA), from 65 to 67. As the FRA increases, the size of the reduction in monthly SSR benefits for those claiming at any age before the FRA declines. There are additional sources of public support for some older workers who experience involuntary earnings reductions. The first is unemployment insurance (UI), which replaces a fraction of lost wages, usually for up to six months, for eligible unemployed workers as long as they are actively seeking work. Specific provisions of this federal- state program vary by state. The most important longer-term benefit available to older workers who experience involuntary wage loss is targeted at those with medical conditions that prevent substantial employment. Social Security Disability Insurance (SSDI) is available to workers with sufficient earnings histories who, upon application, are determined by the Social Security Administration (SSA) to be unable to participate in “substantial gainful activity” (SGA), defined in 2008 for most workers as activity equivalent to generating $940 of labor earnings per month, for a period of at least 12 months.4 SSDI becomes available in the sixth month after the first month in which the worker is determined unable to engage in SGA. 4 The SGA amount for blind beneficiaries is considerably higher: $1,570 in 2008. 2 Employment Support for the Transition to Retirement Supplemental Security Income (SSI) is also available to SSDI beneficiaries whose income (including SSDI) and assets are very low, but SSI benefits are so low that the household income (including benefits) of a large majority of concurrent beneficiaries is below the federal poverty line, and only 16 percent of SSDI worker beneficiaries receive SSI (DeCesaro and Hemeter, 2008).5 SSDI beneficiaries also become eligible for Medicare, but not until the 30th month after the last month in which they earned more than the SGA amount.6 Almost all of those eligible for SSI are also eligible for Medicaid, but in some states the Medicaid means test is more stringent than the SSI means test. The SGA ceiling on earnings creates a strong work disincentive—one that is exacerbated by the five-month waiting period and often lengthy eligibility determination process. Workers must have very low earnings for a lengthy period to obtain benefits. Once on the rolls their benefits will be suspended after their earnings are above the SGA amount for approximately 12 months, and terminated if their earnings continue to be above SGA after an additional 33 months.7 Trade Adjustment Assistance (TAA) and Alternative Trade Adjustment Assistance (ATAA) help much smaller numbers of older workers whose jobs are adversely affected by international competition.8 Others have previously examined the extent to which older workers experience involuntary earnings losses and financial hardship as they approach age 62. Johnson, Mermin and Murphy (2007) examined the extent to which workers age 51 to 55 in 1992, when first interviewed for the Health and Retirement Survey (HRS), experienced a health-related work limitation or a layoff by the age of 62. They found that 25 percent experienced a health-related work limitation and 21 percent experienced a layoff (Table 1). The percentage experiencing a health-related work limitation was significantly higher for females than for males, for those with less than a high school education than for high school or college graduates, and for African Americans than for whites. Those without a college degree were more likely to experience layoffs than those with a college degree, but African Americans were less likely to experience layoffs than whites. Johnson and Mermin (2008) use the HRS to examine the relationship between these two types of adverse events and the likelihood that the worker’s household income falls below the poverty line at age 60 to 61, then again at age 63 to 64, after the worker is eligible for SSR. They report that the household poverty rate for all persons experiencing a health-related work limitation or layoff is 19.0 percent at age 60 to 61, then drops to 5 SSI is also available to low-income working-age people with disabilities who have not had sufficient past earnings to qualify for SSDI. 6 This includes a 5-month SSDI waiting period and a 24-month Medicare waiting period that starts with the first SSDI entitlement month. 7 The text description is a simplification of the SSDI work incentive rules. SSDI has a nine-month Trial Work Period, during which beneficiaries can earn any amount without losing their benefits. The nine months do not have to be consecutive. After the TWP is completed, beneficiaries enter the 36-month Extended Period of Eligibility, during which they lose their benefits in any month in which their earnings are above the SGA amount, with the exception of three-grace period months. Their benefits are terminated after the 36th month if their earnings exceed the SGA amount in the 36th month; if not, full benefits continue until the first month in which earnings exceed the SGA amount, and are then terminated. 8 More information on TAA and ATAA appears in the appendix. 3 Employment Support for the Transition to Retirement Table 1 Incidence of Adverse Employment Events Between 1992 and Year Respondent Turns Age 62, Workers Age 51 to 55 in 1992 Health-related Job Work Limitation Layoff Total 25% 21% Sex Males 23% 21% Females 27%* 21% Marital Status Married 25% 20% Single 27% 23% Education Not high school graduate 32% 23% High school graduate 26%* 21% College graduate 17%* 18% Race and Ethnicity Non-Hispanic white 25% 21% African American 31%* 15%* Hispanic 22% 24% Source: Johnson, Mermin and Murphy (2007), Table 1. Asterisk (*) indicates that the percentage for the given category differs from that of the first category in the group by a statistically significant amount (.05 significance level). 15.2 percent at age 63 to 64. The poverty rate at age 61 to 62 is greater for those with health-related work limitations than for those experiencing a layoff (22.7 versus 12.5 percent), as is the decline in the poverty rate by age 63 to 64 (4.9 versus 3.3 percentage points). They conclude that early entry into SSR plays a substantial role in improving household incomes for both groups. We used SSA administrative statistics along with the findings from the two studies by Johnson and his colleagues to produce rough estimates of the number of fully insured persons age 62 in 2008 who had been working and not on SSDI as they reached age 55 and had experienced a layoff or new health-related work limitation by age 62. We also estimated the number of such workers who had entered SSDI by age 62 and the number with household incomes below the federal poverty line at age 60 to 61 (Table 2). These estimates must be considered very approximate, because they assume that the experiences of those age 62 in 2008 post age 54 were similar to those of HRS respondents who were age 62 in 2000 to 2002, and because of several plausible, but unverified, assumptions, as detailed in the footnotes. Of the approximately 2.4 million fully insured workers who turned age 62 in 2008 and were not already on SSDI, we estimate that 1.1 million had experienced a health-related work limitation or layoff by age 62. Of the 600,000 who had experienced a health-related work limitation, only 150,000 had entered SSDI, and 450,000 had not. Counting those who experienced a layoff, we estimate that 960,000, or 39 percent of those not on SSDI by age 55, experienced at least one of these adverse events and did not obtain SSDI. 4 Employment Support for the Transition to Retirement Table 2 Rough Estimates of Adverse Employment Events Between Age 55 and 62 and Household Poverty at Age 60 to 61 Among Fully Insured Workers Age 62 in 2008 Fully Insured Poverty Prevalence at Age 62 Age 60–61 Percent Thousands Percent Thousands Fully Insured at Age 62 100% 2,600 Entered SSDI by Age 55 7% 170 Not in SSDI by Age 55 93% 2,430 7% 180 No Adverse Event by Age 62 48% 1,320 3% 50 Adverse Event by Age 62 46% 1,110 16% 130 Layoff 21% 510 13% 60 Work-limiting Health Condition 25% 610 15% 120 Entered SSDI by Age 62 6% 150 15% 20 Not in SSDI by Age 62 19% 450 15% 90 Both 7.0% 170 26% 40 Not in SSDI 39% 960 12% 110 Source: Calculations by the author based on Johnson, Mermin and Murphy (2007), Johnson and Mermin (2008) and Social 9 Security administrative data. The adverse events identified in the HRS are not all severe and long lasting, and the households of some workers who experience severe and long lasting adverse events have sufficient income and wealth to maintain their standard of living, despite the adverse event. The poverty statistics indicate, however, that a substantial share of workers who experience one of these events also experience economic hardship by the time they reach age 62. An estimated 130,000 had household incomes below the poverty line at age 60 to 61, including 110,000 who had not entered SSDI. The poverty rate at age 60 to 61 is much lower for those workers who reported neither a layoff nor a health-related work limitation (3 percent versus 16 percent). Because this rate applies to a large number of workers (1.3 million), and because none of these workers entered SSDI, the estimated number with household incomes below the poverty line at age 60 to 61 is substantial, nearly 50,000. We do not know the extent to which these cases reflect the effects of adverse events on the worker’s earnings. 9 The estimate of the number of age 62 fully insured workers in Table 2 is one fifth of the number of fully insured workers age 60 to 64 reported by SSA’s actuaries in December 2007 (http://ssa.gov/OACT/STATS/table4c2FI.html, accessed July 30, 2008). SSA’s actuaries also report that there were 325,666 age 62 SSDI worker beneficiaries in June 2008 (http://ssa.gov/OACT/ ProgData/byage.html?type=da, accessed July 30, 2008). We used SSA’s most recently published death and disability life tables for the 1985 birth cohort of fully insured workers to infer that 53 percent of that number had entered SSDI by age 55 (Baldwin and Chu, 2006). To make the latter calculation we assumed that 10 percent of those on the rolls at age 55 had died or exited the rolls by age 62; the actuarial estimates imply that 1.7 percent of fully insured workers at age 55 are expected to die by age 62. The percentages of insured workers who had experienced an adverse event of each type was inferred from statistics presented in Johnson and Mermin (2008), Table 5; we assumed that all HRS respondents identified as workers by Johnson and Mermin were fully insured and not in SSDI by age 55. We also used the Johnson and Mermin poverty prevalence estimates (Tables 6a and 6b) to infer the prevalence of poverty within each group; we assumed that poverty prevalence within the SSDI and non-SSDI subgroups of those with work-limiting health conditions is the same. 5 Employment Support for the Transition to Retirement We conducted our own analysis of the HRS to determine the extent to which those HRS respondents who were working at first interview reported an onset of a health-related work limitation and experienced a reduction in earnings of at least 25 percent by their last interview before age 62, at age 59 to 61.10 Based on this analysis and SSA program statistics, we estimate that approximately 120,000 fully insured age 62 workers were in this group in 2008. This number is more than 25 percent of the estimated 450,000 workers who experienced the onset of a health-related work limitation after age 54 and did not enter SSDI by age 62 (Table 2). In summary, it appears that: more than one million fully insured workers who had reached age 62 in 2008 had experienced the onset of a health-related work limitation or a layoff after age 54 and before age 62; only about 150,000 of these workers entered SSDI; about 110,000 of those not entering SSDI had household incomes below the poverty line at age 60 to 61; about 120,000 of those with a health-related work limitation had experienced an earnings reduction of at least 25 percent and had not entered SSDI; and the prevalence of poverty declines for these workers by age 63 to 64, once they are eligible for SSR. Although not definitive, these findings suggest that an increase in the EEA would have substantial adverse consequences for the households of a few hundred thousand workers who reach age 62 each year, given current population numbers, plus any additional workers who experience involuntary earnings loss between age 62 and the new EEA. In the remainder of this paper we outline a policy option to address the adverse effects that an increase in the EEA would likely have on the large number of workers who experience involuntary earnings losses as they approach age 62. We call the option Employment Support for the Transition to Retirement (ESTR). We describe eligibility criteria for ESTR in the next section and also consider the likely number of participants under reasonable criteria and discuss the challenges of eligibility determination. In the third section we describe the design of the ESTR benefit itself. In the fourth section, we consider the cost of ESTR in the context of a broader retirement policy reform. We discuss structural issues in the financing and administration of ESTR in the fifth section. We conclude with a comparison of ESTR—in the context of retirement policy reforms— to other approaches for addressing the economic security of workers who experience involuntary earnings losses as they approach age 62, and place ESTR in the broader context of efforts to improve employment security for disadvantaged workers. The antecedents of ESTR can be found in the ATAA, broader efforts to improve benefits for workers who experience long-term job loss late in their careers, and current discussions about policy reforms to address the needs of workers with disabilities. It must be emphasized that the policy option described here is a sketch, intended to encourage discussion of an aspect of retirement policy reform that, in our view, has received inadequate attention. Many different versions of ESTR can be readily imagined, as the discussion in the following sections will illustrate. 10 In the future, we plan to replicate this analysis for layoffs. 6 Employment Support for the Transition to Retirement ELIGIBILITY This section presents and discusses the conceptual definition of the target population, the operational definition of the target population, and options for the eligibility- determination process. TARGET POPULATION CONCEPT In concept, the ESTR target population is workers who experience large involuntary earnings losses as they approach the EEA–losses due to adverse circumstances that are essentially beyond their control. Our assumption is that such workers are the most likely to enter SSR early for lack of a better way to preserve their household’s standard of living. Some will also enter SSDI. We define three types of adverse circumstances. The first is the onset of a significant and long-lasting medical condition that results in a reduction in the worker’s earnings and/or imposes a large financial burden on the household for medical care. The second is a change in the local labor market that substantially reduces the worker’s earnings and earnings prospects. The two most likely examples of the second type of adverse circumstances are (1) a permanent or long-term layoff or wage cut resulting from increased competition, and (2) job loss or earnings reduction in response to a general downturn in the local economy (for example, because of a recession). The third type of adverse circumstance is a substantial increase in the care needs of a dependent or immediate family member with a significant medical condition, thereby substantially increasing household out-of-pocket costs for care and/or imposing time demands on the worker that are incompatible with continued full-time employment at his or her current job. The households of some workers who experience substantial involuntary wage losses will remain economically secure because of other sources of income and we do not exclude such households from eligibility, to avoid imposing a complex means test on eligibility determination. Our expectation, however, is that relatively few such workers would participate because the benefits they would receive under ESTR would be small relative to their household’s income from other sources, and because their own earnings, despite their decline, might remain high relative to the earnings limits on ESTR’s wage subsidies (see the next section). OPERATIONAL DEFINITION OF THE TARGET POPULATION In developing the operational definition of the target population, several major issues require consideration. The first is administrative simplicity. A set of rules that would require extensive data collection, verification, and a complex adjudicative process would burden both the program and the target population. At the same time, however, rules that are very simple will likely mean inclusion of many workers who poorly meet the conceptual definition of the target population and/or exclusion of many who fit it satisfactorily. The second issue is cost. Holding benefit design constant, the less stringent the eligibility criteria, the higher the costs. Costs can also be controlled by the design of benefit features. 7 Employment Support for the Transition to Retirement The third issue is inducement of undesirable behaviors by workers, employers, worker dependents, and perhaps others for the sole purpose of establishing eligibility. Depending on the specifics of eligibility criteria and the eligibility-determination process, undesirable behaviors could include a deliberate reduction in earnings for reasons other than adverse circumstances; increased work in the underground economy (which is to say, work in exchange for payments not reported to the government); reduced savings for retirement; increased waste in expenditures for medical and other services; reduced enrollment in employer-sponsored health plans; transfer of assets to non-dependent relatives or others outside the household; and increased divorce. Benefit design can potentially cause such behaviors, but it can also potentially mitigate the effects of eligibility criteria or the determination process on such behaviors. The fourth issue is the treatment of couples and dependents. In the case of couples, if ESTR’s objective is to enhance family economic security, then eligibility for ESTR could potentially be tied to the percentage reduction in the earnings of both spouses because of adverse circumstances, rather than the percentage reduction in the earnings of the spouse whose earnings were affected. Such a rule would mean that a married worker with given past earnings and an employed spouse would have to experience a larger earnings decline than a single worker with the same earnings to meet a specified percentage reduction criterion—an implicit marriage tax. Such a requirement will be viewed as inequitable by many, which could potentially be addressed imperfectly by a set of rules about how spousal earnings are counted, as well as differences between benefits for married and single workers. Our working assumption is that eligibility will be based on the past earnings of only the worker who has experienced an adverse event, not on the earnings of the worker and his or her spouse. There are other difficult considerations related to spouses or other close family members, however. Some events that cause involuntary wage reductions are events involving spouses and dependents. A family member might require extensive personal care because of an illness or disability, causing the worker to cut back on hours worked or to take a lower paying job that allows the worker to be more attentive to the family member. This is also a difficult area administratively, because the line between voluntary and involuntary reductions in the worker’s earnings is hard to define. A fifth major issue is institutional reactions to ESTR. Employers, financial institutions, insurers, health care providers, and many others interact with the ESTR target population. Assuming that ESTR is a large and well-publicized program, institutional actors will quickly become aware of the program and seek opportunities for gain. Employers, for instance, might try to reduce the wages of some older workers so those workers would qualify for ESTR. The employer would reap some benefits through lower wages and perhaps realize other gains (possibly reduced responsibility for health care costs or subsidies for worksite accommodations). Employers and other institutions have adapted to current employment and retirement policies (Quinn et al. 1990; Burkhauser et al. 1990), and we should expect them to adapt to ESTR as well. We define a structure for operational criteria under which participants must meet two mandatory criteria and at least one of two additional criteria. Operational versions of the mandatory criteria already exist for the administration of Social Security benefits. Certain elements of the additional criteria already have operational versions, but others must be developed. 8 Employment Support for the Transition to Retirement The mandatory requirements are designed to ensure that the benefit is available only to highly experienced workers with a long history of contributions to the Social Security and Medicare trust funds. The requirements are: 1. Social Security earnings history. The worker must have worked long enough to have attained Social Security fully insured status (i.e., the earnings history requirement for retirement benefits).11 2. Minimum and maximum age. The worker must have attained a minimum age (55) and have not attained a maximum age. There are several options for the maximum age. One option would be the EEA, and another the FRA. Alternatively, a time limit on benefit receipt could be imposed (for example, 24 months). The two additional criteria, of which at least one would have to be satisfied, are: 3. Wage loss owing to adverse circumstances. The worker must be a current or former full-time worker who experienced substantial wage loss because of adverse medical, labor market, or family circumstances. To qualify, the worker would have to demonstrate that his or her annual earnings are below a threshold value defined as a percentage of some measure of the worker’s past earnings, for reasons that are beyond his or her control. 4. Existing or former SSDI beneficiary. All existing SSDI beneficiaries meeting the ESTR minimum age requirement would be allowed to exit SSDI and enroll in ESTR. Former SSDI beneficiaries whose benefits were terminated because of earnings and who have not experienced medical improvement would also be eligible. Criterion 2, the age range, has critical implications for both program effectiveness and costs. On the one hand, ESTR eligibility could be limited to those age 62 and older, on the grounds that retirement policy reforms are not likely to have an adverse impact on the target population until they reach age 62, and that costs will be too high if the benefit is offered to younger workers. On the other hand, eligibility at an earlier age, such as 55, would make ESTR available to many workers in the target population when they first experience earnings loss, at the point when support is most likely to result in positive impacts on earnings. A 24-month time limit on eligibility for ESTR could limit the potentially high cost of offering eligibility at age 55. Any time limit could also be applied to just selected components of ESTR benefits (see the next section). 11 Workers must currently be fully insured to be eligible for SSR at age 62. To attain fully insured status, they must have accumulated credits or “quarters of covered earnings” equal to the number of calendar years elapsing between the year in which the worker reaches age 21 and the year in which the worker reaches age 62, becomes disabled, or dies—whichever occurs first. In 2009, one credit was earned for each $1,090 in annual covered earnings up to a maximum of 4 credits for the year. Workers who reach age 62 in 1991 or later need 40 credits to be fully insured. Special rules may apply if workers had an earlier period of disability. The criteria for fully insured status are less stringent than the criteria for “disability insured status,” which includes a “recent work” requirement except for those with whose primary impairment is blindness: at least 20 credits during the 40- calendar-quarter period ending with the quarter of disability onset (SSA 2007a). The earnings history requirement could be made more stringent by requiring disability insured status, which would preclude eligibility for some who are out of the labor force for a substantial period in the years leading up to the minimum eligibility age for ESTR. 9 Employment Support for the Transition to Retirement Criterion 3 is intended to capture workers who have encountered significant involuntary earnings losses after their early fifties. Criterion 4 is intended to entice SSDI recipients back into the labor force and help them increase eventual retirement benefits. Inclusion of those whose SSDI benefits had been previously terminated because of work is intended to avoid discouraging SSDI exit for work before attainment of the ESTR eligibility age. The size of the percentage reduction in earnings and the earnings base used for the percentage reduction in earnings under criterion 3 will be key elements in determining how many workers become eligible for ESTR. One obvious option for the earnings base is to use Social Security’s Average Indexed Monthly Earnings (AIME) value as of a certain age (55, for example). AIME is the measure of past earnings that is used to determine the level of benefits under the current program. Use of this measure would be disadvantageous to workers who were out of the labor force for an extended period, whether voluntarily (to raise children, for example) or involuntarily (illness) relative to a measure that is based on a smaller number of the worker’s highest earnings years. SIZE OF THE TARGET POPULATION AND NUMBER OF PARTICIPANTS As the analysis presented in the first section indicates, a few hundred thousand workers in each single birth-year cohort experience an involuntary earnings reduction and substantial economic hardship as they approach 62 but do not enter SSDI. The number of these workers who become eligible for ESTR will depend on the stringency of the eligibility criteria (that is, what constitutes an involuntary earnings reduction, and how large it must be), and how those criteria are implemented. Given available findings from the HRS, under a reasonable specification, 150,000 to 300,000 workers between the ages of 55 and 62 might become eligible for ESTR in a typical year. If the EEA were increased, some additional number would become eligible after they reach age 62 and before they reach the new EEA. The number of entrants into ESTR each year will presumably be substantially smaller, and will depend on the value of the benefits and time limits. Uptake would be expected to be very low for those who would gain little, and very high for those who would gain much. Given the range of workers becoming eligible each year that is indicated above, an 80 percent uptake rate would mean 125,000 to 250,000 program entrants each year. By comparison, 715,000 workers age 62 were receiving SSR benefits in June 2008, and another 326,000 were on the SSDI rolls.12 The number of participants in any given year will depend on the number of entrants and the duration of participation. If the minimum eligibility age is 55 and entrants at any age can remain in the program until they become eligible for SSR, then the number of participants in any given year will be a large multiple of the number of entrants—five times as many would be plausible. Under that scenario there could be anywhere from 750,000 to 1.6 million participants on the rolls at any time. If, instead, benefit time limits 12 The SSDI figure includes the estimated 120,000 who entered after age 55, as well as those who entered at younger ages. These numbers do not include age 62 dependents of worker beneficiaries unless they are also entitled to Social Security on the basis of their own earnings. Source: SSA, Office of the Actuary, Number of Beneficiaries by Age, June 2008, http://ssa.gov/OACT/ProgData/byage.html?type=ra, Accessed August 3, 2008. 10 Employment Support for the Transition to Retirement result in mean duration of 24 months, the number of participants each year will be approximately twice the number of entrants—perhaps 250,000 to 500,000. Analysis of program costs presented later suggests that the number of participants in a typical year will have to be at the lower end of the figures in the previous paragraph for the program to be fiscally attractive from the government budget perspective. ELIGIBILITY DETERMINATION To the extent feasible, eligibility determination would be based on administrative records, primarily Social Security date-of-birth and earnings records. The administering agency could publish additional information required to document adverse medical, labor market, or family circumstances. Applicants could be required to submit, document, and attest to the veracity of this information, and documentation could include forms to be signed by medical professionals, employers, and perhaps others. Once established, the National Health Information Network would substantially facilitate the process.13 Given the complexity of the information requirements for determining adverse circumstances, the eligibility-determination process would require substantial application assistance that might involve a secure web-based application program to help applicants and their advisors gain access to their Social Security data, determine what additional information is required, assemble all needed information, and submit the application electronically. Such a system would have to be supplemented by paid counselors available to those not capable of navigating the application process alone. Eligibility determination could be largely automated. Intake workers would address problems with administrative records data and conduct simple consistency and validity checks on the information submitted by applicants (to the extent that such checks cannot be fully automated). An automated report would then be generated, indicating the outcome and providing additional information that the applicant and/or agency might need. To minimize administrative burden and maintain short application processing times, program stewardship could potentially rely heavily on a system of random audits and penalties rather than on comprehensive validation of evidence. Applicants could be liable for providing and attesting to false material information (for example, information that affects eligibility or benefit amounts). Medical professionals, employers, and others who certify false material information could likewise be held liable. BENEFIT OPTIONS This section describes and discusses eight potential components of ESTR benefits (Table 3). In turn, each component consists of several options. ESTR would not necessarily include all components described, and even if it did ESTR participants would not necessarily receive all included components, because the size and nature of the 13 Currently, SSA collects most medical records for SSDI applicants from the applicant’s providers by cumbersome and time- consuming methods. SSA is aggressively pursuing electronic collection from providers with well established electronic records systems now, with the expectation that almost all medical records will be collected in this manner once a National Health Information Network is firmly established. 11 Employment Support for the Transition to Retirement Table 3 Summary of ESTR Benefit Components Disability screening and expedited SSDI determinations. Entrants are screened for the likelihood of SSDI eligibility into three groups: (1) those likely to qualify for SSDI and unlikely to engage in substantial gainful activity (SGA) under ESTR, (2) those likely to qualify for SSDI but likely to be able to engage in SGA under ESTR, and (3) those unlikely to qualify for SSDI. Those in the first group could be offered an expedited eligibility determination. Those in the second group would be encouraged or required to try ESTR, with future expedited SSDI determination if they fail after a good-faith effort. Counseling. A counseling system would help participants understand the options available under ESTR, help them address employment issues, and help them develop an economic security plan for the remainder of their lives. Wage subsidy. Those working more than a minimum number of hours per week would be entitled to a time-limited wage subsidy equal to a fraction of the gap between current earnings and an earnings target up to a maximum, based on the participant’s past earnings. Subsidies/allowances for work-related expenses. These would address extraordinary work-related expenses that the worker or worker’s employer must incur. Health insurance subsidy. A subsidy could be made available to participants as well as to employers of participants who offer employee coverage. The subsidy could be in the form of a new Medicare buy-in, an expansion of the Medicaid Buy-in, a refundable tax credit for private health insurance, or publicly funded stop-loss coverage for private plans. Health risk reduction services. Services could be offered to reduce health care risks and improve participant health. Extended unemployment benefits. Unemployed participants could be entitled to additional weeks of unemployment benefits, possibly including incentives to return to work. Supplemental Security Income (SSI) and Medicaid. Those not able to achieve sufficient economic security with other ESTR benefits would be eligible for SSI and Medicaid under current rules for elderly SSI and Medicaid recipients. benefit under each component might depend on such factors as employment, earnings, health care needs, and other income and resources. DISABILITY SCREENING AND EXPEDITED SSDI DETERMINATIONS Offered on a voluntary basis, the first component of ESTR could be an initial screen for ability to participate in SGA and for potential SSDI eligibility. The screening process could be structured to categorize individuals into three groups: (1) those likely to qualify for SSDI and likely unable to engage in SGA under ESTR, (2) those likely to qualify for SSDI but likely able to engage in SGA under ESTR, and (3) those not likely to qualify for SSDI. Individuals in the first group could be offered an expedited SSDI eligibility determination and could potentially be eligible for other ESTR benefits until a final determination is made. Those in the second group could be encouraged or perhaps required to try ESTR first and could be offered the opportunity for future expedited SSDI determination upon meeting conditions designed to ensure a good-faith effort to work, and stay off the SSDI rolls, under ESTR; months spent working under ESTR could potentially be counted toward satisfaction of the SSDI waiting period. Those in the third group and those not requesting an assessment would remain eligible for other components of ESTR. 12 Employment Support for the Transition to Retirement The initial screen is similar to the “triage assessment” proposed by the Social Security Advisory Board (2006) for a redesigned disability program. The screen would serve several purposes. First, it would attract potential SSDI applicants to ESTR and could motivate them to return to or continue work instead of applying for SSDI. Second, the information gathered in the assessment could facilitate the SSDI determination process. Third, those likely to benefit least from ESTR because of a medical condition that prevents SGA under any reasonable circumstances would be quickly moved to a different, more appropriate program. Fourth, those assigned to the second group could potentially be motivated to succeed by satisfying the conditions they would eventually need to fulfill as a condition of obtaining an expedited SSDI eligibility determination. One fiscal concern is that SSDI applicants might routinely enroll in ESTR in order to obtain temporary benefits during the SSDI determination process. Further, SSDI benefits are often awarded retroactively; those who enter SSDI via ESTR could potentially receive income support from both programs for the same period. One way to address potentially overlapping benefits would be to require the transfer of retroactive SSDI benefits to ESTR for the months during which the beneficiary was enrolled in ESTR. As long as protracted delays continue to plague the SSDI determination process, ESTR will remain attractive to SSDI applicants even if retroactive SSDI benefits are transferred to ESTR. BENEFIT, EMPLOYMENT, AND FINANCIAL COUNSELING Many ESTR participants might benefit from counseling on various issues. ESTR involves several components, and participants will need to understand them if they are to use the program to best advantage. Employment counseling is an important component of existing federal employment programs. The target population might also benefit from counseling on retirement finance. Even with ESTR, the economic security of some participants will be relatively low. We expect that many will have limited education and will not have carefully thought through the impact of retirement on household finances. ESTR would likely leverage the capabilities of the existing workforce development system, including the counseling services provided to TAA and ATAA participants. In recent years, One-Stop Career Centers in many states have substantially improved their services for job seekers with disabilities. The centers collaborate with state vocational rehabilitation (VR) agencies and deploy “disability navigators”—staff hired and trained to support One-Stop clients with disabilities. Many large employers offer employees financial planning services for retirement, often through their pension plans. We expect that many ESTR participants would not be eligible for such services and would have to purchase or forgo them. The existing retirement planning industry focuses on clients with relatively high economic security. Hence, developing a financial counseling system to serve ESTR participants might require considerable effort. Not all ESTR participants will need counseling, and some who might need it will likely not use it even if it is readily available without a fee. One option is to require an initial counseling session for all participants to introduce available counselors, convey expectations that participants will use ESTR appropriately, spell out the consequences for failing to do so, and identify participants’ counseling needs. Additional counseling could be 13 Employment Support for the Transition to Retirement optional, although specific types of counseling might be mandatory for some participants. Modest fees could help offset the cost of some types of follow-up counseling. Although counseling might include in-person, one-on-one sessions, a more efficient counseling system could rely on less labor-intensive methods: self-help pamphlets; web- based information and tools, including videos; a program hotline; and group sessions on specific topics. In summary, we envision a counseling system designed specifically for ESTR participants that relies on a variety of counseling methods and draws heavily on currently available counseling systems. WAGE SUBSIDY A wage subsidy is arguably the most important component of ESTR for achieving increased economic security through work. Relative to unemployment benefits, economists view wage subsidies as an efficient way to enhance employment security for those who have experienced wage losses. Several rigorous demonstrations have shown that time-limited wage subsidies, or re-employment bonuses, targeted at dislocated workers result in more rapid return to work and reduced receipt of unemployment benefits (Robins and Spiegelman, eds. 2001). A wage subsidy could be modeled after ATAA’s wage assistance component, which is designed to help the worker attain a target wage level by making up a fraction of the difference between the worker’s actual wage and the target wage. The features of the ESTR wage subsidy might differ from the features of the ATAA subsidy but could potentially include the following: 1. Target wage level. The ATAA target is a measure of previous job earnings up to a maximum of $50,000. 2. Wage replacement rate. This rate is defined as the percentage of the “wage deficit” (that is, target wage minus actual) replaced by the subsidy. The ATAA rate is 50 percent. A higher rate would directly increase program expenditures but would also increase economic security and, by reducing the implicit tax on additional earnings, might also increase earnings and tax revenues. Higher costs associated with a higher replacement rate could potentially be offset by a lower target wage. 3. Maximum subsidy amount. Under ATAA, a maximum subsidy of $10,000 prevents workers from receiving large benefits by accepting jobs that pay extraordinarily low amounts relative to their previous jobs. A maximum subsidy might be appropriate for some ESTR participants, but the adverse circumstances faced by some of them might mitigate against a low blanket maximum. 4. Time limit. Under ATAA, a 24-month limit keeps costs in check and encourages participants to increase their earnings. 5. Employment re-entry time limit. The 6.5-month re-entry time limit under ATAA encourages workers to re-enter the labor force quickly. It might be reasonable to include a re-entry time limit under ESTR, though with a potentially longer time limit for those with medical issues that would prevent early return to work; such a limit could be tied to the time limit for temporary income support (see below). 14 Employment Support for the Transition to Retirement 6. Level-of-effort requirement. The ATAA wage subsidy has a full-time work requirement. That requirement would likely be problematic for workers in the ESTR target population with adverse medical conditions or family circumstances that adversely affect their ability to work full-time. Hence, lower requirements for work effort are appropriate for at least some ESTR participants. A possible effect of a wage subsidy for older workers who have experienced an allegedly involuntary wage loss is that it might induce some to reduce their work effort and partially retire. That might be a desirable outcome for those having difficulty continuing with their current job because of work-limiting health conditions, but will also substantially increase program participation and cost. ESTR eligibility requirements would likely moderate such behavior, but there are practical limits on how effective eligibility requirements can be in this regard. The time limit and level of effort requirement for the subsidy will help restrict the incentive for such behavior. The ATAA wage subsidy has one feature that is not included in the above list: a new employer requirement. The ATAA requirement prevents employers from taking advantage of the ATAA wage subsidy by instituting large across-the-board wage reductions. An employer change requirement would be problematic for ESTR because the most promising opportunities for experienced workers after health-related work limitations are often with their current employers. Safeguards against undesirably opportunistic employer behavior can potentially be built into the eligibility requirements or determination process. The features of the wage subsidy could potentially be tailored to a worker’s specific circumstances. The desirable effects of a tailored subsidy must be weighed against the associated administrative burden, the problems that ESTR participants would face in understanding the subsidy’s implications, and the effort required to provide counseling services. SUBSIDIES/ALLOWANCES FOR EXTRAORDINARY WORK-RELATED EXPENSES The adverse circumstances faced by some ESTR participants might mean that they or their employers face extraordinary expenses if the participants return to work or continue working. Such expenses would likely include: the cost of job accommodations; items that SSDI and SSI currently count as Impairment-Related Work Expenses; expenses for the care during work hours of a dependent with significant disabilities; job-search costs; and relocation expenses associated with a new job. At best, such expenses will reduce the income available to the household for other purposes; at worst, the expenses will make employment impractical. ESTR could include subsidies or allowances for such extraordinary expenses. Under a subsidy system, the subsidy would pay for some portion of such expenses, perhaps with annual and lifetime limits. Participants would submit invoices for specific expenses to obtain approval and reimbursement. A subsidy system would be administratively cumbersome; paperwork, payment delays, and disputes over expense eligibility could also reduce the extent to which the benefit would support ESTR objectives. Further, subsidies for specific goods and services would create incentives for participants and/or employers to purchase more than would be optimal from a social cost- benefit perspective. 15 Employment Support for the Transition to Retirement An allowance program would address some of the limitations of a subsidy program by providing the participant with a fixed allowance to be used at the participant’s discretion, provided the participant works full-time or attains other employment or earnings objectives. The amount of the allowance would be based on some measure of reasonable expenses as well as the participant’s specific circumstances. Allowances would give the participant greater control over purchase decisions and would permit the purchase of necessary supports from informal providers. In addition, allowances would likely reduce the extent to which certain goods and services are purchased in excess of socially optimal levels; that is, the participant would be able to spend the funds on virtually anything, provided that work requirements are satisfied. Such a practice might be viewed as a drawback of an allowance program, especially from a political perspective. One rigorous evaluation of allowance programs for personal assistance services demonstrated positive program outcomes (Brown et al. 2007). In randomized trials, subjects who received cash and counseling to purchase services in three states were significantly and substantially more likely to receive the services they needed and to be satisfied with them than a control group that obtained services paid for directly by an existing program. They were also less likely to experience a long list of adverse events. In some states, the program cost of cash and counseling exceeded the cost of traditional services, but only because reliance on the traditional system often caused control subjects to go without the services they were entitled to and needed. The existing programs were implicitly rationing services through administrative impediments to access. HEALTH INSURANCE SUBSIDIES The nation’s current health care financing system poses an increasing challenge for older workers as the cost of health care escalates and employers continue to eliminate or reduce employer-sponsored benefits. The system is especially problematic for those with chronic medical conditions that might make them eligible for ESTR. The same applies to workers having a dependent with a chronic condition. Workers with a significant medical condition might exit employment earlier than they would otherwise if they lack access to timely health care and/or their condition prevents them from returning to work. At the same time, the cost of employer-subsidized health insurance can motivate employers to terminate employment of workers with chronically high health care expenditures and to not hire prospective employees with such conditions; that might be especially true for self-insured employers, as all medical expenditures paid by employer coverage come directly out of the employer’s bottom line. Further, workers with significant health care needs not met by employer-subsidized coverage have a strong incentive to obtain public coverage either through programs that would make them eligible for Medicaid, or through SSDI, which would eventually make them eligible for Medicare. It is possible that broad reforms to the health care financing system will make health insurance more affordable for older workers and their employers in the near future. For discussion purposes, however, we assume that such reforms will not be implemented any time soon or that such reforms will not adequately address the health care needs of the target population for ESTR. For now at least, health insurance subsidies must be considered an important component of ESTR, although broad health care finance reforms might shape the size and form of the ESTR subsidy. Among the several approaches for subsidizing participant health insurance under our current health care financing system, one would allow participants to buy into Medicare, 16 Employment Support for the Transition to Retirement with employers paying part of the premium commensurate with their contribution to the health insurance of other employees. Such an approach offers several attractive features relative to other options. First, under current law, all ESTR participants become eligible for Medicare at age 65, so this approach would simply advance the age of eligibility for ESTR participants, by up to 10 years (for those who enter ESTR at age 55). Second, the Medicare benefit is already well defined and its administrative infrastructure firmly established. Third, Medicare does not rely on the availability of coverage from a current or past employer, except that, as noted, the employer’s contribution must be commensurate with the employer’s health care expenditures for other workers—zero in some cases. Fourth, a Medicare buy-in would directly reduce the workers’ incentive to obtain SSDI as a means to obtain Medicare even before they reach the age of ESTR eligibility. For example, a 55-year-old worker who meets both SSDI and ESTR eligibility criteria but is just entering the five-month SSDI eligibility waiting period would be able to obtain Medicare coverage via ESTR up to 29 months faster than via SSDI, and without keeping earnings below SGA. They would be able to continue to buy in to Medicare if they stopped working, although the employer would no longer make a contribution. One problem with the Medicare buy-in approach is that individuals might enter ESTR to establish Medicare eligibility and then attempt to enter SSDI as a way of circumventing the Medicare waiting period after SSDI entry. Although termination of Medicare benefits for ESTR participants at SSDI entry could put an end to such behavior, such terminations might be viewed as inequitable and harsh. Perhaps a better option would be to impose a 24-month SSDI waiting period on those who initially enroll in Medicare under ESTR, with waivers for those experiencing significant deterioration of their medical condition after enrollment. In essence, the current Medicare waiting period for SSDI beneficiaries could be replaced by an SSDI waiting period for early Medicare enrollees—a policy that would reward those who continue to work despite their medical condition. The length of the waiting period could be tied to age at ESTR entry–the older at entry, the shorter the waiting period. A second approach would require states to institute a Medicaid Buy-in program for ESTR participants. Many, but not all, states already operate optional Medicaid Buy-in programs for workers with disabilities who meet the medical eligibility criteria for SSDI and SSI benefits. The programs vary substantially in design from state to state. Compared to a Medicare buy-in for ESTR participants, the Medicaid Buy-in approach would likely meet resistance from state governments, result in substantial variation in benefits from state to state, and mean that ESTR participants in many states would be restricted to using providers willing to accept the often stringent Medicaid fee schedule and other restrictive Medicaid features. At the same time, however, some ESTR participants would benefit from Medicaid coverage for services that people with long-term disabilities often need but cannot obtain under Medicare. Following the precedent of the trade adjustment programs, a third approach would provide a subsidy or refundable tax credit for to help pay for coverage. Workers could use the subsidy or credit to help pay for employer coverage, other private coverage if available, or public coverage (i.e., Medicare or Medicaid). To be effective, the subsidy would have to be large enough to assure affordable coverage for low-wage workers. Thus, the tax credit should be refundable and advanceable, and there must be a place where workers without employer-sponsored benefits can go to find affordable and adequate health insurance coverage. An individual tax credit would not provide any 17 Employment Support for the Transition to Retirement medical expenditure relief directly to employers. Such relief might be important for employers of participants with these expenditures, and perhaps it would be feasible to implement the subsidy in a manner that at least partially reduces the employer’s health care expenditures on behalf of the employee. A fourth approach would also rely on access to private coverage: a stop-loss benefit which would limit the liability of the private insurer or, in the case of self-insured employers, the employer. A parallel stop-loss benefit could be made available to participants for out-of-pocket expenditures. In whatever form it takes, the ESTR health insurance subsidy could be a fixed amount for all beneficiaries. Alternatively, a sliding-scale subsidy, based on some measure of household income, would help target the benefit to those most in need, would likely reduce programmatic costs relative to a fixed subsidy, and might hold considerable political appeal, but would require more administrative effort. A time limit on the health insurance subsidy might be especially problematic, as it could mean loss of health insurance for some participants and discontinuity of coverage for others. Use of higher premiums to control the cost of health insurance subsidies would be a better approach than strict time limits. HEALTH RISK REDUCTION SERVICES Many ESTR participants are likely to face the risk of significant health problems that could be addressed through self-care. Such participants could potentially benefit from health risk reduction services designed to improve health and longevity, reduce the demand for high cost health care services, increase productivity, and make it easier to work. Many vendors already deliver inexpensive, largely automated health risk reduction services via the internet, telephone, and mail. These services typically include self- administered health risk assessments; automated, individualized health-risk reduction recommendations generated from information provided by the health risk assessments; referrals to local organizations for assistance in addressing specific health issues and risk factors; and telephone counseling with trained nurses and other health professionals on specific health and health care issues. All of these services are generally designed to augment and support rather than replace services delivered by the individual’s regular providers. The availability and use of health risk reduction services for the working-age population has grown rapidly in recent years. The main purchasers of such services are large corporations and large insurers. Increasingly, private corporations and major insurers recognize that risky behaviors (for example, smoking, poor nutrition, lack of exercise, excessive alcohol consumption, failure to conduct self-examinations, failure to undergo routine checkups and tests, and poor adherence to treatment plans) as well as difficulty in accessing appropriate care impose enormous costs on the health care system and worker productivity. Considerable evidence suggests that well-designed, largely automated, and inexpensive health risk reduction services can have a positive impact on the health of working-age people who take advantage of such services and that the services sometimes pay for 18 Employment Support for the Transition to Retirement themselves over two or three years through reduced utilization of health care services and greater worker productivity, although evidence about their effectiveness as a component of a public program is lacking.14 The Centers for Medicare and Medicaid Services is about to launch a three-year demonstration that will rigorously test the delivery of health risk reduction services to Medicare beneficiaries age 67 to 74.15 If the Senior Risk Reduction Demonstration proves successful, health risk reduction services will likely be available to many seniors through Medicare. The findings from that demonstration might be pivotal to further consideration of health risk reduction under ESTR. Following the approach of Medicare’s Senior Risk Reduction Demonstration, an ESTR health management benefit could purchase services from qualified vendors on behalf of program participants under a performance-based payment system. ESTR participants would be advised of the services at program entry and could potentially be screened to identify individuals most likely to benefit from such services. ESTR design and development could address enrollment incentives, with careful attention to issues that have limited the use of incentives to promote healthy behaviors, including the possibility that incentives might disadvantage participants with significant medical conditions, or encourage such participants to undertake activities detrimental to their health. EXTENDED UNEMPLOYMENT BENEFITS Some—perhaps many—individuals will be unemployed when they enter ESTR, and the adverse circumstances faced by such individuals might pose barriers to a rapid return to work. Many might be eligible for unemployment insurance benefits for up to 26 weeks, but some might not be eligible or might have exhausted much or all of their eligibility before ESTR enrollment. Extended unemployment benefits could be offered to participants who are unemployed when they enter ESTR. The simplest approach—that taken by the trade adjustment programs—would entitle unemployed entrants to a specified number of weeks of benefits beyond those for which they are otherwise eligible. Unemployment benefits, however, reduce the financial incentive to return to work, and the re-employment bonus demonstrations have rigorously shown that duration of unemployment is substantially influenced by the financial incentive to return to work (Robins and Spiegelman, eds., 2001). The negative employment effect of significantly extending benefits to the ESTR target population might be especially large because the length of the extension (for example, 26 weeks) will represent a relatively high share of the period from ESTR entry to SSR eligibility. Hence, a system that selectively provides extended benefits (to those needing an extended period to recover from an illness, for example) might be warranted, even if it proves administratively more complex than a less selective system. Wage subsidies would help offset the work disincentives associated with unemployment benefits. One option that would further reduce the disincentive effect of extended 14 See Aldana (2001), Chapman (1999), Edington et al. (2002), and Fries et al. (1994). Health risk management services are sometimes confused with disease management services but the two are different. Health management services aim to reduce risk and maintain health rather than manage an already existing condition. Disease management is, however, a component of risk reduction for those with chronic illnesses. 15 See http://www.cms.hhs.gov/apps/media/press/release.asp?Counter=1949, accessed March 28, 2008. 19 Employment Support for the Transition to Retirement unemployment benefits would be to count the months of such benefits against the months for the wage subsidy time limit. Some participants will be employed when they enter ESTR but will experience unemployment spells later. Work disincentives associated with extended unemployment benefits for such spells would likely be even more problematic. As the participant approaches SSR eligibility, the length of any extension of unemployment benefits should be reduced. SUPPLEMENTAL SECURITY INCOME AND MEDICAID Currently, SSI and Medicaid are available to almost all working-age individuals who, based on SSA eligibility criteria, are unable to engage in SGA but meet a stringent means test. SSI is also available to all individuals age 65 and over who meet the means test regardless of disability. The current SSI maximum benefit for a single individual is low— approximately 75 percent of the federal poverty level for a single-person household. Some states provide a supplement to the federal benefit, but most do not. In most states, SSI recipients automatically qualify for Medicaid, although the Medicaid means test in some states is more stringent than in others. Some non-SSI recipients qualify for Medicaid under options available in some states, such as the medically needy program. Ideally, ESTR would help participants ensure that their incomes and assets exceed the levels that would make them eligible for SSI and Medicaid, even if they meet SSA eligibility criteria. The extent to which ESTR can achieve such success is unknown, of course, and will depend on the details of program design and implementation. SSI and Medicaid could be offered as a minimal level of economic security for all ESTR participants—regardless of disability—by allowing them to qualify for SSI and Medicaid under the “aged” category. Interactions between the provisions of SSI and Medicaid, on the one hand, and ESTR, on the other, could be problematic. For example, SSI’s work incentive program, Section 1619, reduces benefits by $1 for every $2 of monthly earnings above a small disregard.16 If the ESTR wage replacement rate were set at 50 percent, a working ESTR recipient who also receives an SSI supplement would face a 100 percent benefit reduction rate on marginal earnings. SSI participants and Medicaid enrollees in the target population could potentially be excluded from ESTR on the grounds that the two programs’ benefits overlap in problematic ways. Separation of the two programs would leave those eligible for both with a difficult and complex choice as well as with two completely separate administrative structures. Hence, a better approach might incorporate SSI and Medicaid eligibility into ESTR for those meeting ESTR eligibility criteria and modify the SSI and Medicaid rules for ESTR participants in order to address potential inconsistencies and problematic interactions. Low-income people in this age group receive assistance from other programs, such as food stamps (now the Supplemental Nutrition Assistance Program) and housing 16 The minimum disregard is $85 if the individual has no other income, and $65 if other income exceeds $20. Impairment-related work expenses can add to disregards. 20 Employment Support for the Transition to Retirement assistance. Perhaps special provisions should be added to ensure that low-income ESTR participants are eligible. COST The programmatic cost of ESTR will depend on several factors, including specific details of the benefit design. Our working assumption is that ESTR would be implemented only as a component of a larger retirement policy reform package, and that the overall package would have to produce net savings to the federal government to be politically viable; that is, the cost of ESTR would have to be less than the increases in revenue net of changes in other program expenditures generated by other features of the reform package. It would be premature to fully cost ESTR, but it is possible to illustrate how expensive a program that would substantially address the needs of ESTR’s target population might be, and to compare that expense to the potential savings to the government from a major retirement policy reform, namely a concurrent two-year increase in the EEA. Assume that the EEA is increased to 64 and ESTR is available on a time-limited basis to those age 55 and over who meet the other ESTR eligibility criteria.17 As discussed in Section II.A, under reasonable assumptions about program eligibility, the number of participants could be in the 250,000 to 500,000 range in a typical year even if mean duration is 24 months, and much higher if not. We assume for illustrative purposes that time limits and other limitations on benefits result in a mean duration of 24 months. Program cost per enrollee will depend greatly on specific features of ESTR. Expenditures per SSDI beneficiary, including Medicare expenditures, suggest an upper bound of $28,000 per enrollee per year, but it seems likely that the actual cost per year would have to be well below the cost per SSDI beneficiary to be politically viable.18 For the sake of illustration, assume that the annual cost per enrollee is $15,000. Given the above enrollment range, annual benefit costs would range from $3.75 billion to $7.5 billion per year. Administrative costs are likely to be large relative to program costs, because of the short-term and tailored nature of the program; at 20 percent of program costs (eight times SSA’s estimate for administration of SSDI relative to SSDI benefit costs), they would range from $750 million to $1.5 billion per year, in which case total cost would range from $4.5 billion to $9 billion.19 This range of estimates must be compared to the other likely effects of the two-year EEA increase on benefit costs and federal tax revenues. As illustrated in the next two paragraphs, 17 Presumably a two-year increase in the EEA would be phased in over several years. The analysis focuses on costs once the increase is fully phased in. 18 In December 2006, the mean benefit amount paid to SSDI worker beneficiaries and their spouses and children was $1,503, or $18,045 on an annual basis (SSA, 2007, Table 5.E2). For 2003, estimated Medicare expenditures for under-65 enrollees (almost all of whom are SSDI beneficiaries) not living in institutions were $6,374 (CMS, 2003). We project that the corresponding value for 2007 was approximately $9,600, on the assumption that Medicare expenditures per enrollee in this group increased at the same rate as all Medicare expenditures per enrollee. Expenditures per enrollee in 2003 ($6,966) and 2007 ($10,460) were obtained from Table 1.C1 of Board of Trustees (2004 and 2008, respectively). The large increase from 2003 to 2007 reflects, in part, the introduction of Medicare Part D. The $28,000 figure in the text is the sum of $18,045 and $9,600, rounded. 19 By comparison SSA budgets SSDI administration at 2.5 percent of SSDI benefits (SSA, 2008, Table III.A6). 21 Employment Support for the Transition to Retirement rough calculations suggest that a $4.5 billion program would likely cost less than the savings from the EEA increase, but that a $9 billion program probably would not. SSA would save close to $20 billion in SSR benefits paid to workers age 62 or 63 under this scenario, but our expectation is that all of those savings would eventually be offset by higher benefits at older ages.20 Our assumption is that increases in payroll tax revenues from workers who would claim SSR under current law do not increase their benefits in later years; instead, their benefits are higher when they eventually become eligible only because of the direct effect of later claiming. Whether this assumption is realized will depend on the specific changes to other program rules that interact with the EEA.21 Of more importance to the financing of ESTR is the likely effect of a two-year increase in the EEA on payroll and federal income tax revenue. A $4.5 billion increase in revenue is equivalent to an average increase of $2,700 in payroll and federal income taxes for those age 62 and 63 who would have received SSR in the absence of the EEA increase. That translates into an average increase in earnings of $10,800 ($900 per month) if the average federal tax rate on the additional earnings is 25 percent (including 15.3 percent for both employer and employee payroll taxes and 9.7 percent for federal income taxes). The mean earnings increase required under this scenario is almost as large as the mean monthly benefit amount for age 62 SSR beneficiaries in 2008: $941. A revenue increase of that magnitude or larger certainly seems plausible, although not guaranteed. Under the same assumptions, mean earnings of individuals in this same population would have to increase by at least $21,600 a year ($1,800 per month) to pay for a $9 billion program, a much less likely prospect. Hence, if the overall reform is to produce savings, eligibility rules for ESTR likely need to be tight enough to keep annual enrollment by those who would not be enrolled in SSDI under current law at the lower end of the 250,000 to 500,000 range, or annual program costs per enrollee would have to be lower than the amount we have assumed for this illustration ($15,000) for the cost of ESTR to be less than the savings generated by the increase in the EEA. Reform would affect federal revenues and benefits in additional ways, but other effects would likely be small by comparison. The direction of the effect of the combined reforms on SSDI expenditures and the associated Medicare, SSI and Medicaid expenditures is ambiguous because of the potentially opposing effects of the increase in the EEA and the introduction of ESTR on SSDI enrollment. In the absence of ESTR, an increase in the EEA would no doubt increase enrollment in SSDI. The effect of introducing ESTR at the same time is unclear. Some workers might be induced to enter ESTR rather than SSDI, or be induced to exit SSDI for ESTR, but others might be induced to enter SSDI via ESTR’s expedited SSDI determination process. The direction of the combined effect of an increase in the EEA and introduction of ESTR on TAA/ATAA enrollment and expenditures is also ambiguous, but likely to be very small. 20 In 2007, SSA paid a total of $19 billion in SSR benefits to workers age 62 or 63. (Office of the Actuary, SSA, Number of Beneficiaries by Age, June 2008. http://www.ssa.gov/OACT/ ProgData/byage.html?type=ra. Accessed August 4, 2008.) 21 See SSA (2007a) for details on how the EEA interacts with many other Social Security provisions. 22 Employment Support for the Transition to Retirement FINANCING AND ADMINISTRATION There are multiple options for administering and financing ESTR. Currently several federal and state agencies are responsible for administering various components of ESTR benefits for other populations, and financing comes from a variety of revenue streams, including payroll taxes, general revenues, and employer premiums for unemployment insurance. The key federal agencies include the Department of Labor (DOL) for workforce development and unemployment insurance programs; the SSA for OASDI; the Department of Health and Human Services (HHS) for Medicare and Medicaid; the Department of Education (ED) for vocational rehabilitation (VR) services; and the Department of the Treasury for tax administration. State labor, disability determination, health, and VR agencies play key administrative roles. States also contribute to the funding of their labor, Medicaid, and VR programs, and some states provide income supplements to SSI recipients. ESTR could potentially be implemented through a coordinated set of changes to these agencies’ responsibilities and funding streams at both the federal and state level. Thus, for instance, state labor agencies, under the aegis of DOL, could determine eligibility, with back up support from state disability determination services and SSA for cases involving health-related work limitations; DOL, ED, and their state counterparts could be responsible for various elements of employment and financial counseling services as well as for the subsidies/allowances for extraordinary work expenses, through extensions of workforce development and VR programs; DOL and state labor agencies could administer extended unemployment benefits provided through the unemployment insurance system; wage subsidies could be administered by the federal income tax system; SSA would administer SSDI and SSI related aspects of ESTR; and HHS and its state counterparts would offer health insurance coverage (perhaps as a new Medicare category), take responsibility for risk reduction services, and administer Medicaid benefits for ESTR participants. A system under which the components of ESTR are administered independently by several federal and state agencies is likely to be poorly coordinated, however. Eligible workers would likely find that services are fragmented, making it difficult to obtain what they need and are entitled to. A more desirable model would establish a single point of contact for workers, in local offices run by a state agency under the oversight of a federal agency. The local office, with support from the state agency, would determine eligibility, deliver counseling services, and administer all benefits that are extensions of the agency’s current service programs. Other agencies would be assigned support roles, reflecting their respective roles in administering social service programs. Financing would also come from several sources, again reflecting the funding streams for similar services currently provided to other groups. DOL is the obvious choice for the lead agency, and One-Stop Employment Centers could potentially provide the single point of contact. Even under this model, however, there are likely to be substantial problems related to cross-agency coordination, and individual agencies at both the federal and state level will at times have incentives to work at cross purposes or to siphon funding from ESTR to pay for other agency priorities. The lead federal agency will need to have clear legal authority and the strong backing of both the President and Congress if it is to successfully minimize such behavior. 23 Employment Support for the Transition to Retirement CONCLUSION The primary purpose of this paper is to encourage and expand the discussion of how to address the effects of retirement policy reforms on older workers who experience involuntary wage losses. Past increases in longevity and the expectation of future increases make retirement policy reforms seem inevitable. As of now, 13 of the 29 other countries in the Organization for Economic Cooperation and Development (OECD) have an automatic link between public retirement benefits and expected longevity (Martin, 2008). Compared to the U.S., other OECD countries provide more supports for older workers who experience job loss, and some of them, such as Germany, recently strengthened those supports as they implemented retirement policy reforms (Martin, 2008). From a political perspective, strengthening supports for older workers in this country might be key to moving forward with retirement policy reforms. In the past, much of the domestic discussion on this topic has focused on enhancements to SSDI for older workers. For instance, their claims could be expedited, or eligibility on the basis of vocational factors could be expanded.22 SSDI enhancements would help individuals who experience extreme difficulty in engaging in SGA for medical reasons, but would encourage earlier retirement for others. Those enhancements would not help individuals who experience involuntary wage loss but fail to qualify for SSDI even under the new rules. As seen from our analysis of HRS data, workers who enter SSDI at age 55 or later are a minority of those who experience an adverse event and have household incomes below the poverty line at age 60 to 61, even if we count only those with a health-related work limitation. Expansion of SSDI might be less expensive than many versions of ESTR, but not necessarily all. The most obvious cost of an SSDI expansion would be the difference between SSDI benefits paid under the enhanced rules to those induced to enter SSDI by the policy change and their current law SSR benefits. There would likely also be increases in Medicare costs, as those induced to enter SSDI before age 63 would attain Medicare before age 65, the current Medicare eligibility age for most workers. These direct costs would likely be low relative to the costs of ESTR. What is less obvious is that the revenue gains under a retirement policy reform that includes an expansion of SSDI might be considerably lower than the revenue gains under a reform that includes ESTR instead, but is otherwise the same. The relative costs of the two approaches will depend on specifics of their implementation. The recently proposed Elastic Earliest Eligibility Age is another approach to addressing the potentially adverse effects of a specific retirement policy reform–an increase in the EEA–on disadvantaged workers. One specific proposal for an Elastic EEA would increase the EEA to 64 for those in the top half of the AIME distribution at age 55, but those in the bottom quarter of the distribution would still have an EEA of 62, and those in the second quarter would have an EEA that is somewhere in between, with the exact value depending on the size of the worker’s AIME (Haverstick et al. 2007; Zhivan et al. 22 Under current law, SSA considers the age, education and former occupation (“vocational factors”) of SSDI and SSI applicants age 50 and older, along with medical evidence, in determining the applicant’s ability to engage in SGA. 24 Employment Support for the Transition to Retirement 2008). Another feature of the Elastic EEA is that the benefit level for a worker who claims SSR at the worker’s EEA would be a fixed percentage of what the worker’s benefit level would be if the worker claimed at the FRA, regardless of what the worker’s EEA is. This implies that those with an early EEA would receive a modest increase in benefits relative to current law. The primary cost of the elastic EEA relative to current law is the modest cost of that increase. The authors use HRS data to demonstrate that the prevalence of significant health conditions or disabilities is high among male workers who, under the illustrative Elastic EEA, would have an EEA under age 64.23 This is because those with relatively low life expectancy, poor health, and health conditions that limit work are more likely than others to be in the bottom half of the AIME distribution at age 55. The Elastic EEA is a relatively simple solution to the problem of protecting low-income workers from the adverse effects of an increase in the EEA; it would be easy to administer and involves a relatively simple change to current policy. The target population for the Elastic EEA is different from the target population for ESTR or the target population of an SSDI expansion. The conceptual target of the Elastic EEA is workers with significant health conditions or disabilities, regardless of prior earnings, but the actual target is all workers with low AIMEs, regardless of the cause. ESTR and an SSDI expansion would both target workers who have experienced earnings loss because of an adverse event, with ESTR covering a broader set of adverse events than those covered by an SSDI expansion. The Elastic EEA would not help workers who had high AIME at age 55 but experienced disability onset or another adverse event afterward, and it would help all workers who have low AIME at age 55, including those who do not experience an adverse event as they approach retirement age. As the developers of the Early EEA point out, a large share of those who would remain eligible at age 62 would not have diminished life expectancy, poor health, or health conditions that limit work, and many workers with one or more of these characteristics would not be eligible until age 64. The elastic EEA could be refined to improve targeting, but ultimately targeting will be very imperfect if only earnings are considered. As with ESTR, tighter targeting would require eligibility criteria that are more difficult to administer. Another critical difference between the conceptual underpinnings of ESTR and the Elastic EEA is that the Elastic EEA addresses the income security of its target population through a small increase in income benefits, whereas ESTR does so by providing incentives and support for continued work. Hence, ESTR’s approach is more in line with the purposes of retirement policy reform, and an increase in the EEA that is accompanied by the introduction of ESTR would likely result in larger increases in earnings and tax revenues than the adoption of an Elastic EEA. Not all retirement policy reforms would be as detrimental as an increase in the EEA to workers who experience an involuntary earnings loss as they approach age 62, and a few would benefit some such workers. Elimination of the earnings test for early retirees, reductions in payroll taxes for workers who are paid up, making Medicare the primary 23 The authors to not provide statistics for women. 25 Employment Support for the Transition to Retirement payer for beneficiaries who work, and changing the benefit formula so that benefits are proportional to the number of years with earnings are all examples of proposed reforms that would be favorable to such workers.24 The earnings test has already been eliminated for all SSR beneficiaries older than the FRA, and has increased earnings for those between the FRA and age 70, but also increased the number of workers claiming benefits at or before the FRA (Song and Manchester, 2007a and b).25 Such proposals are not specifically targeted at workers who experience involuntary earnings loss, however, and it seems likely that they would deliver fewer benefits to such workers than a more targeted, intensive program like ESTR. ESTR’s approach is consistent with current thinking about programmatic reforms to address the economic security needs of all working-age people with disabilities. The nation is in the early stages of a “disability policy transition.” We are seeing a shift from policies designed in an era when few people with significant impairments or chronic conditions could support themselves toward policies that reflect advances in medicine and technology and the nature of work such that many people with the same impairments or conditions are able to work.26 ESTR is also broadly consistent with a trend toward public policies that provide assistance to disadvantaged workers through temporary subsidies and other supports that increase the incentive and ability to work. The welfare reforms of the 1990s—the expansion of the Earned Income Tax Credit for low-income parents, time-limited income support, and employment support services—are the most prominent examples.27 The trade adjustment assistance programs, TAA and ATAA, are additional examples. In many respects, ESTR could be viewed as a considerable expansion of ATAA eligibility for workers age 55 and over, to include those who have experienced earnings losses because of adverse events other than increased foreign competition. The approach is the same, but ESTR offers more than ATAA to those with significant health problems or disabilities. Wage subsidies are a key component of both ESTR and ATAA. Brainard et al. (2005, 2006) and LaLonde (2007) have proposed a wage insurance program that would extend the availability of ATAA’s wage subsidy approach to all workers who have been permanently displaced from their jobs and, as a result, face lower earnings prospects. Brainard et al. (2006) point to a Canadian demonstration, conducted in the early 1990s, in which displaced workers received wage supplements of 75 percent of their wage loss for two years up to a maximum of $13,000 per year. The evaluation, which used an experimental design (that is, displaced worker volunteers were randomly assigned to a treatment or control group), found that the program reduced the duration of unemployment by an average 4.4 percent. 24 These proposals are described further in the appendix. 25 The second policy change that has increased the incentive to work longer is elimination of the Social Security earnings test for all SSR beneficiaries older than FRA—the $1 reduction in SSR benefits for every $3 earned above an annual disregard ($15,500 in 1999, the last year the test applied to those over FRA). Before 2000, only those age 70 and older were exempt from the test. Song and Manchester (2007a; 2007b) found evidence, as expected, that elimination of the earnings test increased earnings of those between FRA and age 70 and increased the percentage claiming benefits before age 70. 26 A more extensive discussion of disability policy reforms appears in the appendix. 27 See, for example, Shulz and Hotz (2003). 26 Employment Support for the Transition to Retirement Brainard et al. (2006) estimate that 827,000 workers would have been eligible for their proposed wage subsidy in 2003 and that the cost of a two-year subsidy for those eligible in that year would have ranged from $3.2 billion to $7.0 billion, depending on the share of replaced lost earnings (from 30 to 70 percent in their calculations) and the maximum size of the annual subsidy ($10,000 to $20,000). Their proposed subsidy would presumably be available to all workers in ESTR’s target population, and more. ESTR can be viewed as a more narrowly targeted version of their proposal that also includes additional components intended to support those attempting to work. In 2003, Germany implemented a similar wage subsidy program for older workers. Specifically, job losers age 50 and over are offered two supplements: a wage supplement equal to 50 percent of the gap between previous and current wages and a public pension contribution equal to 90 percent of the gap between previous and new contributions (Martin 2008). The German subsidies have no time limit, and it is possible that they encourage many workers to partially retire. It appears that no follow-up data have been collected, but it would be interesting to establish whether the permanent subsidies have increased or reduced the continuation of full-time work. It must be emphasized that the description of ESTR in this paper leaves open many specifics of the program, and numerous important issues should be further analyzed and debated. We do not yet have sufficient information on the extent to which involuntary wage losses lead to significant financial hardship as workers approach retirement. How should involuntary wage loss be defined? Which workers should be eligible? At what age? For how long? Should eligibility be expanded to include those who have always worked, but never achieved substantial economic success, even if they experience no involuntary wage loss? To what extent should ESTR be a one-stop program serving older workers experiencing a broad array of difficulties, versus two or more separate programs targeted to specific groups of such workers? Along those lines, is it best to approach ESTR as a whole new form of social insurance, a reform to Social Security targeted solely at workers with work-limiting medical conditions, a reform to unemployment insurance targeted at all older workers who experience substantial involuntary wage loss, or something else? What is the right balance between provisions that increase the rewards to work after involuntary wage loss and those that provide the advice and services that workers might need? Should ESTR be largely administered by the federal government, or by state governments with federal oversight? What roles should be assumed by the private sector, including employers, disability management vendors, and private insurers? How can a public program encourage and support private sector efforts to address the needs of the target population without becoming a significant burden to the private sector? How should ESTR be tailored to be consistent with other aspects of retirement policy reforms? These and other issues require much additional examination and debate. Pressing interest in retirement policy reform offers the opportunity to introduce a new employment security approach to a narrowly targeted group of workers, namely those who would otherwise be most adversely affected by the broader reforms. If ESTR is successful and politically popular, it could eventually be expanded to younger workers, and become the only door for SSDI entry. That would be similar to the historical path taken by SSDI itself; when first implemented in 1956, only workers age 50 and older could potentially qualify. For better or worse, introduction of ESTR as a component of retirement policy reform would poke the camel’s nose under the tent. 27 Employment Support for the Transition to Retirement APPENDIX: POLICY BACKGROUND This appendix provides additional information on the policy issues and ideas that motivate ESTR. It includes a summary of the retirement policy reform issue, a summary of current thinking on employment policy reforms for people with disabilities, and a description of the Trade Adjustment Assistance (TAA) and Alternative Trade Adjustment Assistance (ATAA) programs. RETIREMENT POLICY REFORM Several socioeconomic forces have caused policymakers and others to consider policy changes that would encourage later retirement—both a later exit from the workforce and later entry into SSR. The first of these forces is an increase in average longevity and concurrent growth in the percentage of people who remain healthy and able to work well past both the current EEA and even beyond the current FRA. For instance, Cutler, Liebman, and Smyth (2005) report that mortality for men at age 68 in 2000 is equivalent to what it was for men at age 62 in 1960. Steuerle (2008) points out that the average age of retirement was 68 in the decade extending from 1940 to 1950, that the equivalent age in terms of remaining life expectancy in 2005 was 74, and that the equivalent age in 2065 is projected to be 78. Extensive literature documents the decline in disability among those are 65 and older.28 The second socioeconomic force is the aging of the Baby Boom generation. The oldest baby boomers reached the EEA in January 2008. As they continue to age and retire, the number of retirees will grow rapidly, both absolutely and relative to the number of working-age people. As a result, the number of people consuming goods and services will grow rapidly relative to the number producing them, and the number receiving Social Security benefits will grow rapidly relative to the number of people financing Social Security benefits via payroll taxes. The third socioeconomic force is an array of factors causing federal and state expenditures on health care to grow much faster than gross domestic product: rapid growth in the cost of health care per person, the aging of the population, and the availability of Medicare to almost all people age 65 and over. Recent Congressional Budget Office projections of the 75-year outlook under current law imply that growth in expenditures for Medicare, Medicaid, and Social Security will outstrip growth in federal revenues over the next three decades and beyond, eventually leading to enormous deficits unless policies are revamped (CBO, 2007). The argument in favor of policy changes that would encourage later retirement can be stated as follows: If people are healthier, live longer, and therefore consume more goods and services in their lifetimes, it is reasonable to expect them to work during some portion of their additional healthy years, thereby increasing their contribution to the production of goods and services. At the extreme, if people remained healthy and alive forever, we would not expect them to, and could not afford to let them, stop functioning 28 See Schoeni et al. (2008) for a recent review. They attribute improvements in part to declines in heart and circulatory conditions, vision, and musculoskeletal conditions, and in part to advances in assistive and other technologies. 28 Employment Support for the Transition to Retirement as productive members of society at age 65. Policies that delay retirement would help address projected budget deficits by helping increase GDP and federal revenues and reducing expenditures for Social Security and, possibly, Medicare. Although the above argument is compelling, it ignores the fact that gains in longevity and healthy years are not distributed uniformly across all workers. In fact, gains in longevity for men since the early 1990s have been concentrated among those in the top half of the earnings distribution during their prime work years. Waldron (2007) examined Social Security earnings histories and longevity for 29 successive birth cohorts of men from 1912 to 1941. She classified men who were fully insured for Social Security at age 60 into the top and bottom half of the earnings distribution on the basis of their earnings between age 45 and 55, excluding those with no earnings during that period. For each birth cohort, she used Social Security mortality data to determine average years of life from age 60 to 89 for those in each half of the earnings distribution. For the most recent cohort, she found that those in the top half of the earnings distribution lived, on average, 5.8 years longer than those in the bottom half. For the first cohort this difference was only 1.2 years. She found that average remaining years of life at age 60 for those in the top half of the earnings distribution increased from 18.9 for the 1912 cohort to 25.4 for the 1941 cohort—a gain of 6.5 years. In contrast, the increase for those in the bottom half of the earnings distribution was only 1.9 years, from 17.7 to 19.6. Several earlier studies cited by Waldron have documented large and widening longevity differences for men across levels of education. National statistics show significant differences in life expectancy at age 65 by race.29 Two recent policy changes are already increasing the incentive to work longer. The first is a gradual increase in FRA from age 65 to age 67, starting with those turning age 62 in 2000 and ending with those turning age 62 in 2031. Workers may still obtain early retirement benefits at age 62, but the value of benefits for those claiming at age 62 is gradually declining because the Social Security Administration (SSA) has reduced benefits for those claiming at age 62, making the benefits actuarially equivalent to those claimed when workers attain FRA. For those turning age 62 in 2000, the benefit claimed at age 62 was 80 percent of the benefit claimed at FRA; the percentage will decline to 70 percent for those turning age 62 in 2031. Song and Manchester (2007b) have shown that the change has led to a gradual reduction in claims for Social Security benefits between age 62 and FRA. No evidence has yet emerged about effects on earnings. Many analysts are advocating additional policy changes to encourage later retirement. In January 2008, the Social Security Advisory Board convened a forum to discuss policy changes that would encourage later retirement. These include: 1. Educational campaign. An educational campaign would advise people about their likely financial needs in later years and encourage them to work longer and claim benefits later (Munnell 2008; Steuerle 2008). In contrast, current messages from the government appear to encourage early claiming, such as SSA’s publicity surrounding the early claiming of the first baby boomer to reach age 62. Other examples include a clearer explanation of how benefits increase with later claiming. 29 Life expectancy for white males at age 65 is 17.2 years compared to 15.2 years for African American males (National Center for Health Statistics 2007). 29 Employment Support for the Transition to Retirement 2. Administrative barriers to early claiming. Current administrative barriers to claiming Social Security at age 62 or later are minimal; SSA makes the process as effortless as possible for the claimant. One specific policy suggestion would require the consent of the early claimant’s spouse, if any, to ensure that the spouse understands the implications of early claiming for spouse survivor benefits. As the spouse is often a woman and women often outlive their husbands, such a policy change might reduce the number of elderly women who find themselves in financial straits after the husband’s death. 3. Increase the EEA. An increase in the EEA is the policy change most frequently advocated as a way to encourage later retirement. One version of the change would gradually increase the EEA to 64 for those reaching age 64 in 2033. Given that the same cohort would have an FRA of 67, three years would separate the EEA and FRA, just as before the recent increase in FRA. The benefit for those who claim at the EEA would once again equal 80 percent of the benefit for those who claim at FRA. Another policy change would eliminate the EEA. 4. Increase the Medicare Eligibility Age. The Medicare Eligibility Age (MEA) has been 65 since the program’s inception. One proposal would increase it to FRA. 5. Additional increases in FRA. As many have pointed out, since the 1935 inception of Social Security, the scheduled two-year increase in FRA has been well below the average increase in life expectancy for those who reach age 62. One option is to index FRA to mortality rates; another option is to index FRA to expected years of life remaining. The EEA could also be indexed (Butrica, Smith, and Steuerle 2006). Martin (2008) reports that 13 countries in the Organization for Economic Co- operation and Development (OECD) have in some way already indexed public pension plans to life expectancy (Australia, Demark, Finland, France, Germany, Hungary, Italy, Mexico, Norway, Poland, Portugal, Slovak Republic, and Sweden). 6. Change the Social Security benefit formula. Several suggestions for changing the method of benefit computation would increase the incentive to work longer and claim later: a. Increase the number of earnings years counted in determining Average Indexed Monthly Earnings (AIME), which is the basis of the Primary Insurance Amount PIA (that is, the worker’s benefit amount, not counting dependent benefits), from 35 to 40 (Goda et al. 2008). Currently, years of earnings after 35 years at best replace earlier years with lower earnings so that the impact of additional years of earnings on the benefit amount can be small. b. Disentangle benefit progressivity from career length. Under this proposal, AIME would be based only on years with positive earnings. The progressive benefit formula would be applied to this newly defined AIME, just as it is now, but the result would be multiplied by years with positive earnings divided by 35 (Goda et al. 2008). This change would substantially increase the effect of working longer on retirement benefits, especially after 35 years. c. Remove the earnings test for early retirees, likely increasing early retirement but also likely increasing earnings among those who retire early. 30 Employment Support for the Transition to Retirement d. Back load benefits; that is, substantially reduce the benefit at age 62 and increase it as the beneficiary ages up to a specified age by, for example, using the wage index rather than the price index to make annual benefit adjustments (Steuerle 2008). e. Impose a maximum on the benefits of early retirees; those with high earnings contributions would not be able to claim full benefits until FRA. 7. Reduce or eliminate payroll taxes for those who continue to work after a specified number of years. Such a policy would increase older workers’ incentive to work and motivate employers to retain older workers. Goda et al. (2008) suggest a category of “paid up” workers—those with substantial earnings for at least 40 years. The worker’s PIA would then be fixed. Participation in the paid up program could be at the worker’s option, to allow for that fact that some workers might prefer to continue replacing low-earnings years (in their top 35 earnings years) with high-earnings years, thereby continuing to increase their PIA. 8. Make Medicare the primary payer for beneficiaries with employer-based health insurance. Current rules make Medicare the secondary payer if a beneficiary works and has access to employer-based coverage. The result is a disincentive to work. The burden of health care expenditures falls on the worker and employer if the beneficiary continues to work but falls on Medicare if the worker retires. Some of these proposals, if implemented by themselves, would be problematic for older workers in adverse circumstances. Increases in the various eligibility ages and back- loaded benefits are especially problematic. Some of the policies, however, would benefit at least some workers with low economic security: eliminating the earnings test for early retirees, reducing or eliminating payroll taxes for those who have paid up, making Medicare the primary payer for all beneficiaries, and changing the benefit formula that would make benefits proportional to the number of years with earnings. DISABILITY POLICY REFORM The United States has started what appears likely to be a slow and often difficult transition from disability policies that came of age in the mid-20th century to the promulgation of policies that are “modern” in the sense that they reflect the medical, technological, workplace, and social realities of the 21st century (Stapleton et al. 2006). Although the current mix of public and private supports has served millions of workers well, many workers still fall between the cracks while many others leave the workforce prematurely and join the permanent disability transfer rolls. Compared to other households, the households of working-age people with disabilities report much lower average size-adjusted incomes, and many more live in poverty. Further, the best available evidence indicates that these gaps are growing. At the same time, government expenditures to support working-age people with disabilities are increasing at a rapid rate.30 30 Goodman and Stapleton (2007) estimate that in 2002 the federal government spent $226 billion to support working-age people with disabilities and that states spent an additional $50 billion, primarily for Medicaid. Federal expenditures represented 11.3 percent of all federal outlays, up from 6.1 percent in 1984. 31 Employment Support for the Transition to Retirement Policymakers, program administrators, and advocates have already made significant efforts to strengthen the public support system for workers with disabilities. The policies of greatest relevance for our purposes are those designed to provide support for workers after they experience the onset of a medical condition that limits their ability to work. The most important program for such individuals is SSDI, which is the gateway to the second most important program, Medicare. SSDI is often described as a “medical retirement” program in that it provides Social Security benefits to workers deemed no longer able to engage in SGA because of a medically determinable condition. The SGA earnings level ($940 per month in 2008) is just over the federal poverty level for a family of one and somewhat less than an individual can earn by working 40 hours per week at the federal minimum wage.31 Beneficiaries become eligible for Medicare after 24 months of SSDI eligibility. In 2002, SSDI and Medicare expenditures for working-age people with disabilities totaled $108 billion (Goodman and Stapleton 2007). Once workers enter SSDI, they are unlikely to exit the program for a reason other than age (for example, SSR entry) or death, especially if they enter after age 50, as most do (Rupp and Scott 1998). Federal and state governments provide some support for workers attempting to continue working or to return to work after disability onset, but the available support is meager in comparison to that provided to those who medically retire via SSDI. Most notably, state vocational rehabilitation (VR) agencies, largely financed by the federal government under the Rehabilitation Act, provide services to help workers return to work. Total federal expenditures on VR services are extremely small compared to SSDI and Medicare expenditures—just $2.5 billion in 2002. Public support for medical care, which can be critical to continuation of or return to work, is particularly limited for those who have never entered SSDI or SSI, and varies by state. Although SSDI and Medicare successfully serve millions of workers who experience disability onset, the programs are less successful in serving those who, with either a modest level of support or perhaps even temporary support, might continue working or return to work. VR experts stress that such workers need to receive support before they separate from their employer or at least before they experience a prolonged separation from the labor force and become dependent on SSDI or other non-labor income. The SSDI determination process encourages them to do just the opposite. By law, such workers must demonstrate that they cannot engage in SGA by earning less than the SGA level for at least five months.32 In practice, the period during which they must earn less than SGA is often much longer because of delays in the disability determination process, especially if benefits are awarded after an initial denial, as is often the case. About 70 percent of benefit awards are made at the initial determination level, another 6 percent after a review of the initial denial, and the remaining 24 percent only after an appeal to SSA’s Office of Hearings and Appeals.33 In FY 2007, the mean processing time for initial eligibility determinations was 83 days, and the mean processing time for appeals was 512 31 This is the Social Security Administration’s SGA earnings level for most beneficiaries in 2007. A higher value of $1,500 applies to blind beneficiaries. The poverty guideline for a family of one is $817 per month. The federal minimum wage in 2006 was $5.15 per hour; at that wage rate, monthly earnings based on a 40-hour work week would total $893. Automatic cost-of-living adjustments apply to each of these amounts except the minimum wage. Source: Social Security Administration, 2007. 32 See Social Security Advisory Board (2003) for a discussion of whether the statutory definition of disability should be changed. 33 Social Security Advisory Board (2006b), Chart 69. 32 Employment Support for the Transition to Retirement days (SSA 2007b). To obtain a benefit award, the applicant’s earnings must not exceed the SGA level during the entire period, and the applicant must not engage in SGA. The incentive to keep earnings below the SGA amount continues once beneficiaries are on the rolls because beneficiaries risk loss of benefits if they earn above SGA for more than 12 months. The incentive is especially strong during the 24-month Medicare waiting period, which starts in the first month after SSDI entitlement.34 Both policymakers and advocates have demonstrated considerable interest in developing and implementing an “early intervention” program designed to support continuation of work or return to work after disability onset. In theory at least, a well-designed program could provide workers with an option preferable to SSDI and Medicare that is less costly for the government. At the same time, SSDI and Medicare would continue to operate for the many who truly cannot contribute substantially to their own support. In a position paper supported by several leading advocacy organizations, MacDonald and O’Neil (2006) have called for a new social insurance program, funded by payroll taxes, called Employment Support Insurance (ESI). Similarly, the Social Security Advisory Board (2006) has called for the development and testing of a public support system under which workers experiencing disability onset would be assessed and then directed to either SSDI or an alternative program designed to help them return to work. In its final report to the president and Congress, the Ticket to Work and Work Incentives Advisory Panel (2007) called for multi-agency demonstrations for policies that would encourage employers to retain workers after disability onset and provide them with short-term assistance. The National Council on Disability (2007) also calls for a series of multi-agency employment demonstrations. The Council of State Administrators of Vocational Rehabilitation (CSAVR) has proposed testing an Employment First program under which state VR agencies would offer temporary support and return-to-work services to SSDI applicants as an alternative to SSDI entry (CSAVR, 2008). All of these proposals revolve around sets of supports designed to encourage and help the worker’s household maintain its standard of living during the period immediately after disability onset and to help the worker return quickly to a high level of self-sufficiency through work. The supports include: 1. Assessment and expert counseling 2. Expedited entry into SSDI for those deemed unable to attain a high level of self- sufficiency through work 3. Financing for health care to ensure early access to appropriate medical care 4. Financing for supports and services to help the individual continue working or return to work 5. Incentives and assistance for the employer to retain the worker 34 The entitlement month often precedes the SSDI award month because retroactive awards are made for past months. If SSDI benefits are terminated because of work during the Medicare waiting period, the beneficiary will fail to obtain Medicare coverage and will be unable to take advantage of (1) the eight-year Medicare Extended Period of Eligibility, which is available to those who exit the rolls for work after they obtain Medicare, and (2) eligibility for continuation of Medicare enrollment after that period via payment of a premium, provided that the medical condition that was the basis for SSDI continues (SSA 2008). 33 Employment Support for the Transition to Retirement 6. Wage subsidies or partial income support after return to work, perhaps for a limited period 7. Short-term income support for the period before return to work One approach to implementation of such a program would increase employer incentives to retain workers after disability onset and, at the same time, leverage the return-to-work services that private disability insurers and disability management vendors deliver on behalf of employers and their employees (Stapleton et al., forthcoming). TRADE ADJUSTMENT ASSISTANCE In 1962, the Kennedy Administration established the Trade Adjustment Assistance (TAA) program to compensate workers displaced as a result of either increased imports within their industry or shifts to overseas production associated with trade liberalization. The program, administered by the U.S. Department of Labor (DOL) with the support of state workforce development programs, provides a variety of benefits and re-employment services to all members of worker groups that have been certified as eligible in response to a group petition. In 2002, Congress overhauled the program by streamlining the application process, expanding eligibility, adding a health coverage tax credit, doubling the training budget, and substantially increasing budget outlays for income support. Recognizing that retraining might be less appropriate for older workers, Congress also established the parallel ATAA program, which offers a wage subsidy to eligible workers at least 50 years of age who accept lower-wage re-employment. After a worker group is certified, each worker may apply for individual services through DOL’s local One-Stop Career Centers to determine TAA or ATAA eligibility. The eligibility process is such that only those workers who meet TAA criteria are evaluated for ATAA eligibility; in the case of dual eligibility, workers must elect to receive benefits under one of the two programs (Table A.1). Both programs provide a host of re-employment services and a tax credit for 65 percent of monthly health insurance premiums (Table A.2). Reflecting their disparate target worker populations, the programs differ chiefly in the design of cash benefits. TAA provides job-search allowances (for seeking work outside the worker’s region of residence), relocation allowances, and trade readjustment allowances (a 26-week extension of unemployment insurance benefits) to provide support during the transition to new employment. Given that older workers are less likely to invest in relocation or training, ATAA replaces support for those activities with a wage subsidy that is designed to encourage a rapid return to employment, even if the return necessitates a wage reduction. The subsidy is equal to 50 percent of the difference between the old and new wages up to $10,000 for a two-year period and is available only after attainment of a new full-time job at an annual wage of less than $50,000 within 6.5 months of separation from the former job. In 2006, 120,000 workers were eligible for TAA, and program expenditures totaled $259 million. Most recipients received Trade Readjustment Allowances, and many enrolled in training. We do not know how many received wage subsidies under ATAA, but we understand that the number represents only a small share of those eligible for TAA. With ATAA explicitly designed for older workers, it is of interest as a model for ESTR. Unfortunately, only limited information is available on the performance of this new 34 Employment Support for the Transition to Retirement Table A.1 Eligibility Criteria for Trade Adjustment Assistance Trade Adjustment Assistance (TAA) 1. Workers’ company produces a product (not a service); and 2. A required minimum of the workforce has been laid off in the 12 months preceding the date of the petition or is threatened with layoffs (3 workers in groups of fewer than 50, or 5 percent of the workforce in groups of 50 or more); and At least one of the following must apply: 1. Increased imports contributed importantly to an actual decline in sales or production and to a layoff or threat of a layoff; or 2. There has been a shift in production to certain countries outside the United States; or 3. There has been a shift in production outside the United States, and there has been or is likely to be an increase in the import of like or similar articles; or 4. Loss of business as a supplier of component parts, a final assembler, or a finisher for a TAA-certified firm contributed importantly to an actual decline in sales or production and to a layoff or threat of a layoff. Alternative Trade Adjustment Assistance (ATAA) The worker group must meet the TAA criteria and the following: 1. A required minimum of the petitioning workers are age 50 or older (3 workers in groups of fewer than 50, or 5 percent of those in groups of 50 or more); 2. Workers age 50 or older in the work group possess skills not easily transferable; and 3. Conditions within the industry are adverse. Source: Department of Labor website at http://www.doleta.gov/tradeact/benefits.cfm#1. program. Brainard et al. (2005) characterized ATAA as slow to be implemented and noted that its eligibility criteria are poorly defined within a larger, excessively complex process (TAA). Two years after enactment of the 2002 TAA reform legislation, the nation had only 1,115 ATAA participants. Brainard et al. also cited problems in restricting eligibility to workers who “lack easily transferable skills” but are nonetheless able to achieve re-employment within 26 weeks, especially given that this standard applies in addition to the TAA eligibility rules. The authors of the report suggest a broadened definition of eligibility that would include any full-time worker who experiences a substantial reduction in earnings to a level below a specified maximum—in essence, insurance against substantial wage loss despite full-time work. 35 Employment Support for the Transition to Retirement Table A.2 Summary of TAA and ATAA Benefits Benefits TAA ATAA Rapid Response Assistance   Employment Counseling   Resume and Interview Skills Workshops   Career Assessment   Job Development   Job-Search Programs   Job Referrals   Job-Search Allowances  Relocation Allowance  Job Training  Trade Readjustment Allowances (TRA)  Wage Subsidy  Health Coverage Tax Credit   Source: U.S. Department of Labor at http://www.doleta.gov/tradeact/benefits.cfm#1. 36 Employment Support for the Transition to Retirement REFERENCES Aldana, S. G. “Financial Impact of Health Promotion Programs: A Comprehensive Review of the Literature.” American Journal of Health Promotion 15 (5): 296–320, 2001. Baldwin, R., and S. Chu. A Death and Disability Life Table for Insured Workers Born in 1985. Actuarial Note Number 2005.6, Social Security Administration, Baltimore, MD: SSA, February 2006. Brainard,L., R. E. Litan, and N. Warren. “A Fairer Deal for America’s Workers in a New Era of Offshoring.” In L. Brainard and S. M. Collins, eds., Offshoring White-Collar Work—The Issues and Implications. Washington, DC: The Brookings Institution, 2006. Brainard, L., R. E. Litan, and N. Warren. “Insuring American Workers in a New Era of Offshoring.” Policy Brief 143. Washington, DC: The Brookings Institution, 2005. Brown, R., B. L. Carlson, S. Dale, L. Foster, B. Phillips, and J. Schore. “Cash and Counseling: Improving the Lives of Medicaid Beneficiaries Who Need Personal Care or Home- and Community-Based Services.” Princeton, NJ: Mathematica Policy Research, August 2007. Burkhauser, R. V., and J. F. Quinn. “Economic Incentives and the Labor Force Participation of Older Workers. Research in Labor Economics 11:159–180, 1990. Butrica, B. A., K. E. Smith, and C. E. Steuerle. “Working for a Good Retirement.” Annandale on Hudson, NY: The Levy Economics Institute Working Paper Series No. 463, July 2006 Chapman, L. S. “Proof Positive: An Analysis of the Cost-Effectiveness of Worksite Wellness.” Seattle, WA: Summex Corporation, 1999. Congressional Budget Office. Long-term Budget Outlook. Washington, DC: CBO, December 2007. Council of State Administrators of Vocational Rehabilitation. Employment First. Washington, DC: CSAVR, 2008. Cutler, D. M., J. B. Liebman, and S. Smyth. “How Fast Should the Social Security Retirement Age Rise?” Preliminary Draft. Cambridge, MA: Harvard University and the National Bureau of Economic Research, April 25, 2005. DeCesaro, A. and J. Hemeter. “Characteristics of Non-institutionalized DI and SSI Program Participants.” Office of Policy Research and Statistics Note No. 2008-02, Baltimore, MD: Social Security Administration, 2008. Edington, M., T. Karjalainen, D. Hirschland, and D. W. Edington. “UAW-GM Health Promotion Program: Successful Outcomes.” AAOHN Journal 50 (1): 26–31, 2002. Fries, J. F., H. Harrington, R. Edwards, L.A. Kent, and N. Richardson. “Randomized Controlled Trial of Cost Reductions from a Health Education Program: The California Public Employees’ Retirement System (PERS) Study.” American Journal of Health Promotion 8 (3): 216–223, 1994. 37 Employment Support for the Transition to Retirement Goda, G. S., J. B. Shoven, and S. N. Slavov. “Removing the Disincentives for Long Careers in the Social Security and Medicare Benefit Structure.” Paper prepared for the Social Security Advisory Board’s Conference “Working Toward Retirement Security: Policies to Extend the Working Life of Older Americans,” Washington, DC, January 18, 2008. Goodman, N., and D.C. Stapleton. “Federal Expenditures for Working-age People with Disabilities.” Journal of Disability Policy Studies 18 (2): 66–78, 2007. Johnson, R.W., G. B. T. Mermin, and D. Murphy. “The Impact of Late-Career Health and Employment Shocks on Social Security and Other Wealth.” Retirement Project Discussion Paper 07-07. Washington, DC: The Urban Institute: 2007. Johnson, R. W., and G. B. T. Mermin. “Economic Hardship Before Qualifying for Social Security Benefits.” Working Paper. Washington, DC: 2008. Haverstick, K., M. Sapozhnikov, R. Triest, and N. Xhivan. “A New Approach to Raising Social Security’s Earliest Eligibility Age.” Working Paper 2007-19. Chestnut Hill, MA: Center for Retirement Research at Boston College, 2007. Lalonde, R.J. “The Case for Wage Insurance.” CSR No. 30. New York: Council on Foreign Relations, September, 2007. MacDonald, B., and M. O’Neil. Being American: The Way Out of Poverty. Oakland, CA: World Institute on Disability, 2006. Martin, J. P. “Some Suggestions for Reforms Based on OECD Countries’ Experiences.” Office of Employment, Labour and Social Affairs, Organization for Economic Co-operation and Development, Paris, France: OECD, January 18, 2008. Munnell, A. “Changing Employment Incentives and Claiming Behavior.” Presentation to the Social Security Advisory Board’s Conference, “Working Toward Retirement Security: Policies to Extend the Working Life of Older Americans.” Washington, DC, January 18, 2008. National Center for Health Statistics. Health, United States, 2006 with Chartbook on Trends in the Health of Americans. Hyattsville, MD: NCES, 2007. National Council on Disability. Empowerment for Americans with Disabilities. Breaking Barriers to Careers and Full Employment. Washington, DC: NCD, 2007 Quinn, J. F., R. V. Burkhauser, and D.A. Myers. Passing the Torch: The Influence of Economic Incentives on Work and Retirement. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 1990. Robins, P. K. and R. G. Spiegelman (editors). Reemployment Bonuses in the Unemployment Insurance System: Evidence from Three Field Experiments, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 2001. Rupp, K., and C. G. Scott. “Determinants of Duration on the Disability Rolls and Program Trends. In K. Rupp and D.C. Stapleton, eds., Growth in Disability Benefits: Explanations and Policy Implications. Kalamazoo, MI: W. E. Upjohn Institute for Employment Research, 1998. Schoeni, R. F., V. A. Freedman, and L. G. Martin. “Why Is Late-Life Disability Declining? The Milbank Quarterly 86 (1): 47–89, 2008. 38 Employment Support for the Transition to Retirement Shulz, K. M., and V. J. Hotz. “The Earned Income Tax Credit.” In R. Moffitt, ed., Means-Tested Transfer Programs in the United States. Chicago, IL: The University of Chicago Press; Cambridge, MA: National Bureau of Economic Research, 2003. Social Security Administration. The Red Book—A Guide to Work Incentives. Baltimore, MD: SSA, 2008. Social Security Administration. Annual Statistical Supplement, 2006. Baltimore, MD: SSA, 2007a. Social Security Administration. SSA’s Performance and Accountability Report for Fiscal Year (FY) 2007. Baltimore, MD: SSA, 2007b. Social Security Administration. Office of Policy. Considerations for Potential Proposals to Increase the Earliest Eligibility Age. Policy Brief 2007-01. Baltimore, MD: SSA, 2007c. Social Security Administration. The 2008 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds. Baltimore, MD: SSA, 2008. Social Security Advisory Board. The Social Security Definition of Disability: Is it Consistent with a National Goal of Supporting Maximum Self Sufficiency? Washington, DC: SSAB, 2003. Social Security Advisory Board. A Disability System for the 21st Century. Washington, DC: SSAB, 2006a. Social Security Advisory Board. Disability Decision Making: Data and Materials. Washington, DC: SSAB, 2006b. Song, J. G., and J. Manchester. “New evidence on earnings and benefit claims following changes in the retirement earnings test in 2000.” Journal of Public Economics 91(3–4): 669–700, 2007a. Song, J. G., and J. Manchester. “Have People Delayed Claiming Retirement Benefits? Responses to Changes in Social Security Rules.” Social Security Bulletin 67 (1): 1–23, 2007b. Stapleton, D. C., R.V. Burkhauser, P. She, R. Weathers II, and G. A. Livermore. “Income Security for Older Workers.” Journal of Disability Policy Studies, forthcoming. Stapleton, D. C., B. O’Day, G. A. Livermore, and A. J. Imparato. “Dismantling the Poverty Trap: Disability Policy for the 21st Century. The Milbank Quarterly 84 (4):701–732, 2006. Steuerle, C. E. “Improving Financial Security Through Longer Working Lives.” Presentation to the Social Security Advisory Board’s Conference, “Working Toward Retirement Security: Policies to Extend the Working Life of Older Americans,” Washington, DC, January 18, 2008. Ticket to Work and Work Incentives Advisory Panel. Building on the Ticket: A New Paradigm for Investing in Economic Self-Sufficiency for People with Significant Disabilities. Final Report to the President and Congress. Washington DC: SSA, 2007 Waldron, H. “Trends in Mortality Differentials and Life Expectancy for Male Social Security–Covered Workers, by Average Relative Earnings.” Social Security Bulletin 67(3):1–18, 2007. Zhivan, N., S. Sass, M. Sapozhnikov, and K. Haverstick. “The Elastic EEA: A New Approach to Raising Social Security’s Earliest Eligibility Age.” Issue in Brief. Chestnut Hill, MA: Center for Retirement Research at Boston College, 2008. 39