This is the accessible text file for GAO report number GAO-14-311 entitled 'Retirement Security: Challenges for Those Claiming Social Security Benefits Early and New Health Coverage Options' which was released on May 27, 2014. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. United States Government Accountability Office: GAO: Report to the Chairman, Subcommittee on Primary Health and Aging, Committee on Health, Education, Labor, and Pensions, U.S. Senate: April 2014: Retirement Security: Challenges for Those Claiming Social Security Benefits Early and New Health Coverage Options: GAO-14-311: GAO Highlights: Highlights of GAO-14-311, a report to the Chairman, Subcommittee on Primary Health and Aging, Committee on Health, Education, Labor, and Pensions, U.S. Senate. Why GAO Did This Study: Deciding when to retire and claim Social Security benefits can be one of the most important financial decisions older Americans make. Despite higher monthly benefits for those who delay, many people still claim Social Security retirement benefits at age 62, the earliest age of eligibility. In 2014, these early claimers will see their monthly benefits reduced by 25 percent compared to what they would have received if they had delayed claiming until age 66, the current full retirement age. At the same time, some early claimers do not have access to government or employer-sponsored health insurance. These early claimers may have been able to purchase coverage on the individual market, but they may have also been subject to denials and rate increases because of their health status. To better understand the circumstances faced by those who claim early Social Security benefits, GAO examined: (1) demographic and occupational characteristics associated with early claiming; (2) retirement income of early claimers compared to those who delay; and (3) how PPACA changes health coverage options for early claimers. More specifically, GAO examined the characteristics and income of early claimers using data from the Health and Retirement Study, as well as the eligibility for PPACA insurance programs using 2009-2011 American Community Survey data. GAO received technical comments on a draft of this report from the Department of Labor and the Social Security Administration and incorporated them as appropriate. What GAO Found: Several work-related factors may cause people to claim Social Security benefits early and suggest they may face challenges in continuing to work at older ages. For example, those who worked in physically- demanding blue collar jobs were 55 percent more likely to claim benefits prior to their full retirement age compared to those in all other occupations after controlling for other factors (see figure). Those who were out of the workforce or had longer work histories by age 60-62 were also significantly more likely to claim early. In addition to work-related characteristics, other factors, such as having poorer expectations of living to age 75 significantly increase the likelihood of claiming early. Figure: Work-Related Factors Affecting the Likelihood of Claiming Benefits Early: [Refer to PDF for image: horizontal bar graph] Worked at least 35 years by age 60-62: More likely: 38%[A]. Blue collar job at age 60-62: More likely: 55%[A]. Less than a college degree: More likely: 23%[B]. Employed full-time at age 60-62: Less likely: 30%[A]. Managerial or professional job at age 60-62: Less likely: 32%[A]. Source: GAO analysis of Health and Retirement Study data. [A] indicates that this variable is statistically significant at the 99 percent level. [B] indicates that this variable is statistically significant at the 90 percent level. [End of figure] Those who delay claiming until their full retirement age tend to have greater income and wealth in retirement and rely less on Social Security than those who claim earlier. Our analysis shows that the median income for those who delay was 45 percent higher after claiming benefits than for those who claimed early, and 33 percent higher at age 72. Although delayed claimers have higher median Social Security benefits, those benefits make up a smaller portion of household income than for early claimers. Even when comparing early and delayed claimers with similar total income after claiming, average household income for delayed claimers was higher at age 72 than for early claimers. However, for both early and delayed claimers, Social Security benefits accounted for an increasing share of total income as they aged. In 2014, some early claimers, especially those without access to health coverage, may benefit from certain provisions of the Patient Protection and Affordable Care Act (PPACA) intended to improve the availability and affordability of health coverage. GAO estimates that nearly a million early claimers did not have government or employer- sponsored health insurance before 2014. Of these, 14 percent may be newly eligible for Medicaid in 2014 due to expansion in 25 states and the District of Columbia and 58 percent could be eligible for tax credits that reduce the premiums for coverage purchased through the new health insurance exchanges. However, GAO estimates that 10 percent of these early claimers had incomes below the federal poverty level but lived in states that did not expand Medicaid and had incomes too low for federal exchange tax credits. View [hyperlink, http://www.gao.gov/products/GAO-14-311]. For more information, contact John Dicken, (202) 512-7114, dickenj@gao.gov or Charles Jeszeck, (202) 512-7215, jeszeckc@gao.gov. [End of section] Contents: Letter: Background: Work-Related Factors and Perceptions of Life Expectancy Affect Decision to Claim Social Security Benefits Early: Those Who Delay Claiming Social Security Have Somewhat Higher Income in Retirement, Even among Those with Similar Income at the Start of Retirement: Nearly A Quarter of Early Claimers Lacked Access to Government or Employer-Sponsored Health Coverage and Thus May Benefit From New Options, Mostly at Reduced Costs: Concluding Observations: Agency Comments: Appendix I: Objectives, Scope, and Methods: Appendix II:Average Household Income and Wealth for Households near 10th and 90th Percentile of Total Household Income: Appendix III:Estimates of the Number and Percentage of Early Claimers Potentially Eligible for Expanded Medicaid in 2014: Appendix IV:Comments from the Social Security Administration: Appendix V:GAO Contacts and Staff Acknowledgments: Tables: Table 1: PPACA-Defined Income Categories Used to Calculate Premium Tax Credit Amounts for Health Insurance Purchased through the Exchanges: Table 2: Example of Benefit Differences under the Cost-Sharing Subsidies: Table 3: Sample Statistics: Table 4: Factors Associated with Claiming Social Security Benefits Prior to the Full Retirement Age: Table 5: Estimated Early Claimer Eligibility for Medicaid in Those States Approved to Expand Medicaid, by State: Figures: Figure 1: Adjustments to Monthly Benefit Based on Claiming Age Where Full Retirement Age Is 66: Figure 2: Share of Eligible Men and Women Claiming Social Security Benefits by Age and Birth Cohort: Figure 3: Share of Those Claiming Benefits at Age 62 or Any Time Prior to Full Retirement Age by Occupation, 2000-2010: Figure 4: Several Work-Related Factors Affect the Likelihood of Claiming Benefits Early after Controlling for Other Factors: Figure 5: Share of Those Claiming Social Security Benefits Early by Various Demographic Characteristics: Figure 6: Share of Those Claiming Social Security Benefits Early by Total Household Wealth Quartile: Figure 7: Share of Those Claiming Early by Expected Probability of Living to Age 75 and Total Household Wealth Quartile: Figure 8: Median Household Income and Wealth Statistics, by Claiming Age, after Claiming Benefits (in 2010 Dollars), 1992 to 2010: Figure 9: Household Income Statistics, by Social Security Claiming Age and Income Source, After Claiming Benefits, 1992 to 2010 (Annual, in 2010 Dollars): Figure 10: Median Household Income and Wealth Statistics, by Social Security Claiming Age, at Age 72, 1992 to 2010 (in 2010 Dollars): Figure 11: Household Income Statistics, by Social Security Claiming Age and Income Source, at Age 72, 1992 to 2010 (Annual, in 2010 Dollars, Except as Noted): Figure 12: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72, 1992 to 2010 (in 2010 Dollars)--Households near Median Income: Figure 13: Early Claimers without Government or Employer-Sponsored Coverage Who May Be Eligible for Expanded Medicaid or Federal Tax Credits for Coverage Purchased through the Exchanges: Figure 14: Estimated Early Claimer Health Insurance Coverage in the United States, 2009 through 2011: Figure 15: Average Monthly Premiums Paid by a 62-Year-Old Nonsmoker for the Second-Lowest-Cost Silver Plan in Each State: Figure 16: Monthly Premium Paid by a 62-Year-Old Nonsmoker after Applicable Premium Tax Credits for the Second-Lowest-Cost Silver Plan, by Income Level: Figure 17: Average Percent of Income Paid to Premiums for a 62-Year- Old Nonsmoker after Applicable Premium Tax Credits for the Second- Lowest-Cost Silver Plan, by Income Level: Figure 18: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72 (in 2010 Dollars)-- Households near 10th Percentile Total Household Income: Figure 19: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72 (in 2010 Dollars)-- Households near 90th Percentile Total Household Income: Abbreviations: ACS: American Community Survey: CMS: Centers for Medicare & Medicaid Services: COBRA: Consolidated Omnibus Budget Reconciliation Act of 1985: FPL: federal poverty level: HHS: Department of Health and Human Services: HRS: Health and Retirement Study: MAGI: modified adjusted gross income: PPACA: Patient Protection and Affordable Care Act: SSU: second stage sampling of area segments: SSA: Social Security Administration: VA: Department of Veterans Affairs: [End of section] United States Government Accountability Office: GAO: 441 G St. N.W. Washington, DC 20548: April 23, 2014: The Honorable Bernard Sanders: Chairman: Subcommittee on Primary Health and Aging: Committee on Health, Education, Labor, and Pensions: United States Senate: Dear Mr. Chairman: The age at which someone chooses to retire and claim Social Security benefits can be among the most important financial decisions American retirees make. Eligible workers can claim Social Security retirement benefits as early as age 62, but the monthly benefit is lower for the rest of a retiree's life than if they delayed claiming.[Footnote 1] The Social Security benefit formula adjusts monthly payments so that someone living to average life expectancy should receive about the same amount of benefits over their lifetime regardless of which age they claim. While claiming early provides retirement income at an earlier age, the reduction in benefits may cause those retirees to regret the decision if they live a long time. For 62-year-olds who claim in 2014, monthly benefits will be 25 percent lower than they would be if claimed at age 66, the current full retirement age--a reduction that will grow to 30 percent as the full retirement age increases to 67 by 2022. Increases in average number of years spent in retirement and the decline of defined benefit pensions that provide lifetime income only increase the importance of the Social Security claiming age decision.[Footnote 2] Retiring before Medicare eligibility at age 65 may also create the challenge of maintaining affordable health insurance early in retirement.[Footnote 3] Some early claimers may qualify for health insurance through a government-sponsored program or a former employer, though the percentage of employers offering such coverage has been declining.[Footnote 4] Those early claimers without access to these sources of coverage may be able to purchase coverage directly from insurers through the individual market. However, older individuals have historically been more likely to be denied coverage or charged higher premiums. The Patient Protection and Affordable Care Act (PPACA) offers new health insurance options for such individuals, in particular for those with low incomes.[Footnote 5] For example, states may opt to expand their Medicaid programs to cover a broader category of low-income adults,[Footnote 6] and beginning in 2014, individuals can purchase coverage through health insurance exchanges in each state and may also qualify for federal assistance to make that coverage more affordable. The availability of these new options could provide those who decide to claim early Social Security benefits with new sources of affordable health care. Given the financial consequences of claiming Social Security benefits before the full retirement age, you asked us to explore the issues facing Americans who claim early with a special focus on their occupational characteristics, the impact on retirement income, and the recent changes related to PPACA. Specifically, this report examines the following questions: 1. What demographic and occupational characteristics are associated with claiming Social Security benefits before the full retirement age? 2. How does the composition and distribution of retirement income for those who claim Social Security retirement benefits early compare to those who delay? 3. To what extent did early claimers lack health insurance prior to the recent implementation of PPACA provisions, and how have their coverage options since changed? To determine which factors are associated with claiming Social Security benefits prior to the full retirement age, we reviewed relevant literature and interviewed experts and agency officials to discuss their research methodologies and findings. Based on this review, we analyzed data from the Health and Retirement Study (HRS), primarily from 2000 to 2010. Using HRS data, we examined key characteristics of early and delayed claimers, such as demographics, health status, and work history. For purposes of this report, we define early claimers as those who report receiving Social Security benefits at any point between age 62 and 1 month prior to their full retirement age. We also examined those who claimed at age 62, the earliest age of eligibility, and present descriptive statistics for this sub-population of early claimers in parts of our report. To focus our analysis on claiming retirement benefits, we exclude those who report receiving Social Security disability income from our sample. [Footnote 7] We also applied other exclusion criteria as described in appendix I. We analyzed descriptive statistics and developed a standard statistical model to determine the significance of these factors. We also compared the level and source of income and wealth, at the time of claiming and later in retirement, for retirees who claim Social Security at different ages. Unless otherwise noted, all estimates presented in this report have a relative standard error of less than 15 percent. To examine the extent to which early claimers lacked health insurance prior to the recent implementation of PPACA provisions, and how their coverage options have since changed, we examined 2009-2011 data from the American Community Survey (ACS) of the U.S. Census Bureau. [Footnote 8] We examined health insurance coverage for early claimers from age 62 through 64 who are not yet eligible for Medicare based on their age. Among these early claimers, we identified the group of individuals who were either uninsured or had an individual plan purchased directly from a health insurance issuer as their only source of coverage. We used the 2009-2011 data to estimate the total number of early claimers who met income eligibility standards who would be eligible for expanded Medicaid or for the premium tax credits and cost- sharing subsidies for plans offered through the exchanges.[Footnote 9] We also obtained data from the Department of Health and Human Services (HHS) and the Henry J. Kaiser Family Foundation on premiums for health plans offered through the exchanges.[Footnote 10] In each state, we identified premiums for a nonsmoker age 62 and, for comparison, age 21. To assure the premiums we report represent what a large number of individuals in the state would pay, we selected the premiums offered in the most populous county. To provide background data on the premiums available through the individual market prior to 2014, we examined January 2013 base premiums for individuals age 19 and 64 in the individual market that we previously reported.[Footnote 11] To describe select PPACA provisions that could impact health insurance options for early claimers, we reviewed relevant federal laws, regulations and guidance and other documents issued by HHS and the Internal Revenue Service.[Footnote 12] We examined the reliability of HRS, ACS, and the premium data by interviewing the appropriate officials, reviewing documentation, and conducting selected data checks. Based on these assessments, we determined that the data were reliable for our purposes. (A more detailed description of our scope and methodology is provided in appendix I.) We conducted this performance audit from May 2013 through April 2014 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. Background: Claiming Social Security Benefits: Social Security retirement benefits cover 96 percent of American workers, as well as their spouses and survivors. Monthly retirement benefits depend on the beneficiary's work and earnings history and the age at which the beneficiary chooses to begin receiving benefits. [Footnote 13] Eligible workers may elect to begin receiving Social Security retirement benefits as early as age 62; an age-62 claimer, however, will receive a reduction in monthly benefits compared to what she would receive if she waited to claim until reaching her full retirement age.[Footnote 14] The full retirement age has risen from 65 originally to 66 for many current new retirees to 67 for workers born in 1960 or later.[Footnote 15] As the full retirement age has risen, so has the relative reduction in benefits when claimed at age 62 or any time prior to full retirement age. An eligible worker claiming at age 62 in 2014, with a full retirement age of 66, would have his monthly benefit reduced by 25 percent (see figure 1).[Footnote 16] An age-62 claimer starting in 2022, with a full retirement age of 67, would receive monthly benefits 30 percent below what he would receive had he claimed at 67 (and 44 percent below what he would receive monthly starting at age 70).[Footnote 17] As we have previously reported, the availability of Social Security benefits at age 62 and adjustments to benefits thereafter may give workers mixed signals about whether to retire early or continue working.[Footnote 18] Figure 1: Adjustments to Monthly Benefit Based on Claiming Age Where Full Retirement Age Is 66: [Refer to PDF for image: vertical bar graph] Recipient's claiming age: 62; Percentage of Primary Insurance Amount (PIA): 75%. Recipient's claiming age: 63; Percentage of Primary Insurance Amount (PIA): 80%. Recipient's claiming age: 64; Percentage of Primary Insurance Amount (PIA): 87%. Recipient's claiming age: 65; Percentage of Primary Insurance Amount (PIA): 93%. Recipient's claiming age: 66; Percentage of Primary Insurance Amount (PIA): 100%. Recipient's claiming age: 67; Percentage of Primary Insurance Amount (PIA): 108%. Recipient's claiming age: 68; Percentage of Primary Insurance Amount (PIA): 116%. Recipient's claiming age: 69; Percentage of Primary Insurance Amount (PIA): 124%. Recipient's claiming age: 70; Percentage of Primary Insurance Amount (PIA): 132%. Source: GAO analysis of Social Security Administration documentation. Note: The Primary Insurance Amount is the benefit a person would receive if he/she elects to begin receiving Social Security retirement benefits at his/her full retirement age. [End of figure] According to the Social Security Administration (SSA), someone who lives to an average life expectancy should receive about the same amount in lifetime benefits regardless of claiming anytime between age 62 and 70. However, several factors can affect the tradeoffs of claiming Social Security at different ages. For example, how long an eligible worker expects to live--based on family history or current health status--can affect when a worker claims. For married couples, the life expectancy of both spouses affects the lifetime household benefits received because Social Security pays survivor benefits to widowed spouses.[Footnote 19] Other factors affecting someone's claiming decision include other sources of income, including earned income, pension coverage, a spouse's income and health, or psychological factors.[Footnote 20] Given the complexity of this decision-making process, we have previously reported on the importance of SSA statements as a financial literacy tool that can educate the public about SSA program benefits and aid in financial planning. [Footnote 21] Increasing life expectancy and time spent in retirement raise the potential cost to retirees of claiming early, and therefore reduced, benefits. According to data from the Centers for Disease Control and Prevention, life expectancy at age 65 has increased from 16.7 years in 1983, when statutory changes raised the future full retirement age, to 19.1 years in 2010. However, gains in longevity have not been shared equally across the population. In particular, life expectancy at 65 for white males has risen more over that period than for other groups, such as black men. Further, a 2012 study indicated that Americans with less education have made few gains in life expectancy since the 1950s and 1960s.[Footnote 22] And women with less education have actually seen a decline in life expectancy.[Footnote 23] Nonetheless, several studies that have examined the Social Security benefit formula and claiming decision process show that many people stand to gain from delaying claiming benefits.[Footnote 24] The share of men and women claiming Social Security benefits at age 62 has diminished in recent years. As shown in figure 2, of those eligible for retired worker benefits born in 1935 who reached the full retirement age of 65, about 43 percent of men and 49 percent of women claimed benefits in the first month after turning age 62. For those born in 1946, for whom the full retirement age is 66, the share claiming in the first month after turning age 62 declined considerably to about 32 percent for men and 38 percent for women. Still, 62 remains the most prevalent age to claim Social Security benefits, and the large majority of eligible workers claim by their full retirement age. At the same time, a relatively large share of men and women choose to claim at their full retirement age--a spike that has tracked with the phase-in of the increase in the full retirement age from 65 to 66 (see figure 2). This pattern is noteworthy given that the age of eligibility for Medicare has remained at 65. In total, for those eligible for Social Security benefits born in 1946, few delayed claiming 1 year or more past their full retirement age (only about 8 percent of men and 7 percent of women claimed at age 67 or later). Figure 2: Share of Eligible Men and Women Claiming Social Security Benefits by Age and Birth Cohort: [Refer to PDF for image: 2 multiple line graphs] Female retired workers: Age at claiming: 62; Women born in 1935: 49%; Women born in 1940: 43%; Women born in 1946: 38%. Age at claiming: 63; Women born in 1935: 1%; Women born in 1940: 1%; Women born in 1946: 1%. Age at claiming: 64; Women born in 1935: 1%; Women born in 1940: 1%; Women born in 1946: 1%. [Phase-in of increase in Full Retirement Age (FRA)] Age at claiming: 65; Women born in 1935: 14%; Women born in 1940: 2%; Women born in 1946: 2%. Age at claiming: 66; Women born in 1935: 0; Women born in 1940: 0; Women born in 1946: 13%. Age at claiming: 67; Women born in 1935: 0; Women born in 1940: 0; Women born in 1946: 0. Age at claiming: 68; Women born in 1935: 0; Women born in 1940: 0; Women born in 1946: 0. Age at claiming: 69; Women born in 1935: 0; Women born in 1940: 0; Women born in 1946: 0. Age at claiming: 70; Women born in 1935: 1%; Women born in 1940: 2%; Women born in 1946: 0. Male retired workers: Age at claiming: 62; Men born in 1935: 43%; Men born in 1940: 38%; Men born in 1946: 32%. Age at claiming: 63; Men born in 1935: 1%; Men born in 1940: 1%; Men born in 1946: 1%. Age at claiming: 64; Men born in 1935: 1%; Men born in 1940: 1%; Men born in 1946: 1%. [Phase-in of increase in Full Retirement Age (FRA)] Age at claiming: 65; Men born in 1935: 20%; Men born in 1940: 15%; Men born in 1946: 2%. Age at claiming: 66; Men born in 1935: 0; Men born in 1940: 0; Men born in 1946: 18%. Age at claiming: 67; Men born in 1935: 0; Men born in 1940: 0; Men born in 1946: 0. Age at claiming: 68; Men born in 1935: 0; Men born in 1940: 0; Men born in 1946: 0. Age at claiming: 69; Men born in 1935: 0; Men born in 1940: 0; Men born in 1946: 0. Age at claiming: 70; Men born in 1935: 0.1%; Men born in 1940: 0.5%; Men born in 1946: 0. Source: GAO analysis of data from the Social Security Administration, Office of the Chief Actuary. [End of figure] Health Coverage Options for Early Claimers Prior to 2014: Early claimers' eligibility for health insurance is dependent on their age, employment, income, and other factors. Those individuals age 65 and older are eligible for coverage through Medicare. Those individuals younger than 65 may be eligible for an employer-sponsored plan[Footnote 25] or for government-sponsored programs, such as Medicare,[Footnote 26] Medicaid, Department of Defense programs, or Department of Veterans Affairs (VA) programs.[Footnote 27] Individuals not eligible for these sources of coverage may choose to buy coverage directly through the individual market. Prior to PPACA, issuers offering individual health insurance plans generally were not subject to uniform federal requirements and therefore protections available to consumers varied amongst states. For example, the majority of states did not impose a guaranteed issue requirement for issuers of individual plans, which would prohibit issuers from denying enrollment of individuals because of their health status. In addition, many states did not restrict issuers in the individual market from adjusting health insurance premiums based on an individual's health status.[Footnote 28] Older individuals seeking coverage through the individual market have historically been more likely to be denied coverage or charged higher premiums based on such factors as their health status or age. For example, America's Health Insurance Plans reported in 2009 that the denial rate progressively increased as the applicant's age increased, from a low of 5 percent for applicants under 18 years of age to a high of 29 percent for applicants from 60 to 64 years of age.[Footnote 29] One reason for this is that older individuals are more likely to have preexisting conditions.[Footnote 30] We recently reported on base premium amounts prior to underwriting for coverage offered in the individual markets of all states as of January 2013.[Footnote 31] These data also show variation in premiums due to location, gender, and age. For example, median base premiums for a 64- year-old male nonsmoker were 4.2 times the premiums for a 19-year-old, on average. In five states that we previously identified as having protections similar to PPACA prior to 2014,[Footnote 32] premiums varied little, or did not vary, by age; in other states, median base premiums for older individuals were more than 5 times those of younger individuals. Furthermore, because these are base premiums, the average premium for a 64-year-old after underwriting may be higher relative to a 19-year-old, because older individuals are more likely to have preexisting conditions that would be assessed during underwriting. Select PPACA Provisions Related to Coverage: Several PPACA provisions that are effective as of January 2014 provide additional protections to individuals obtaining coverage through the individual market.[Footnote 33] For example, PPACA requires issuers offering individual plans to guarantee coverage and prohibits them from basing eligibility or coverage on health status, medical conditions, and other such factors.[Footnote 34] The act also prohibits issuers from charging higher premiums based on health status, the use of health services, or gender. In addition, plans are prohibited from varying premiums between young and old individuals by more than 3 times and tobacco users cannot be charged more than 1.5 times the premiums charged to those who do not use tobacco.[Footnote 35] Therefore, under PPACA, a plan's premiums for an individual within the same geographic area cannot vary by more than 4.5 times. PPACA also requires that individual plans available through the exchanges beginning in 2014 must cover certain categories of benefits and provide standardized coverage levels. Specifically, plans are categorized as bronze, silver, gold, or platinum depending on the portion of health care costs expected to be paid by the health plan. [Footnote 36] For example, a bronze plan will typically have lower premiums and higher cost-sharing (such as deductibles and copayments) when services are delivered, while a platinum plan will have higher premiums and lower cost-sharing. PPACA also made important changes to the way eligibility for the Medicaid program is determined and who the program covers. Unlike prior federal Medicaid eligibility rules that required an individual to meet both income and categorical requirements--such as low-income children, pregnant women, and certain individuals with disabilities-- PPACA allows states to cover most adults who have incomes at or below 133 percent of the federal poverty level (FPL) under their State Plans.[Footnote 37] As of December 2013, 26 states had been approved by the Centers for Medicare & Medicaid Services (CMS) to expand Medicaid in 2014.[Footnote 38] PPACA also created health insurance exchanges, which are health insurance marketplaces intended to allow individuals to compare, select, and enroll in health coverage from among participating carriers. PPACA directed each state to establish an exchange by January 1, 2014.[Footnote 39] In states electing not to establish and operate such an exchange, PPACA requires the federal government to establish and operate an exchange in the state.[Footnote 40] In 2014, the federal government established and is operating the exchange in 34 states. Individuals purchasing insurance through the exchanges may also be eligible for income-based premium tax credits and cost-sharing subsidies to reduce the cost of that insurance. PPACA includes a federal tax credit to lower the premium cost for eligible individuals to between 2 and 9.5 percent of income for individuals whose incomes are from 100 to 400 percent FPL. (See table 1.) To be eligible, individuals may not have other qualified coverage, such as through a government-sponsored program or their employer.[Footnote 41] In addition, PPACA provides cost-sharing subsidies for certain eligible individuals from 100 to 250 percent FPL to lower the out-of-pocket costs--deductibles and copayments--that are paid when an individual receives care.[Footnote 42] Table 1: PPACA-Defined Income Categories Used to Calculate Premium Tax Credit Amounts for Health Insurance Purchased through the Exchanges: Income categories: Percentage of the federal poverty level (FPL): At least 100%, but less than 133% FPL; Applicable income[A]: $11,490 to $15,282; Maximum percentage of income to be paid to premiums: Lower bound: 2.0%; Upper bound: 2.0%. Income categories: Percentage of the federal poverty level (FPL): At least 133% FPL, but less than 150% FPL; Applicable income[A]: $15,282 to $17,235; Maximum percentage of income to be paid to premiums: Lower bound: 3.0%; Upper bound: 4.0%. Income categories: Percentage of the federal poverty level (FPL): At least 150% FPL, but less than 200% FPL; Applicable income[A]: $17,235 to $22,980; Maximum percentage of income to be paid to premiums: Lower bound: 4.0%; Upper bound: 6.3%. Income categories: Percentage of the federal poverty level (FPL): At least 200% FPL, but less than 250% FPL; Applicable income[A]: $22,980 to $28,725; Maximum percentage of income to be paid to premiums: Lower bound: 6.3%; Upper bound: 8.05%. Income categories: Percentage of the federal poverty level (FPL): At least 250% FPL, but less than 300% FPL; Applicable income[A]: $28,725 to $34,470; Maximum percentage of income to be paid to premiums: Lower bound: 8.05%; Upper bound: 9.5%. Income categories: Percentage of the federal poverty level (FPL): At least 300% FPL, but not more than 400% FPL; Applicable income[A]: $34,470 to $45,960; Maximum percentage of income to be paid to premiums: Lower bound: 9.5%; Upper bound: 9.5%. Source: GAO analysis of Internal Revenue Service and Department of Health and Human Services (HHS) documents. [A] This table presents household income as a percentage of FPL in 2013 for individuals living in the 48 contiguous states and the District of Columbia. The FPL for households with additional persons or that are located in Alaska or Hawaii are higher. [End of table] Work-Related Factors and Perceptions of Life Expectancy Affect Decision to Claim Social Security Benefits Early: Holding a Blue Collar Job and Other Work-Related Factors May Cause People to Claim Social Security Benefits Early: Several work-related factors may cause people to claim early and suggest they may face challenges in continuing to work until older ages. In particular, our analysis of HRS data shows that the type of job a worker holds will affect whether that worker will claim Social Security benefits prior to their full retirement age. Compared to all other occupations, those who held a blue-collar job at age 60 through 62 were 55 percent more likely to claim early, after controlling for other factors.[Footnote 43] These occupations also tend to be more physically demanding. For example, 61 percent of those who held a blue- collar job at age 60 through 62 reported their job was physically demanding and/or involved heavy lifting most or all of the time compared to 25 percent of those in all other occupations. As shown in figure 3, the share of people claiming early varies widely when comparing more detailed categories of occupation. A much smaller share of those in managerial and professional occupations, which tend to be less physically demanding, claim benefits prior to their full retirement age. These differences by occupation also hold when only looking at those who claimed at age 62. Figure 3: Share of Those Claiming Benefits at Age 62 or Any Time Prior to Full Retirement Age by Occupation, 2000-2010: [Refer to PDF for image: 2 horizontal bar graphs] Percentage claiming Social Security benefits: Men: Managerial: At age 62: 26%; Before full retirement age: 56%. Professional: At age 62: 22%; Before full retirement age: 51%. Sales: At age 62: 36%; Before full retirement age: 66%. Clerical/Administrative: At age 62: 40%; Before full retirement age: 62%. Services: At age 62: 45%; Before full retirement age: 69%. Health services: At age 62: N/A; Before full retirement age: N/A. Farming: At age 62: 54%[A]; Before full retirement age: 81%. Mechanics: At age 62: 45%; Before full retirement age: 80%. Construction: At age 62: 49%; Before full retirement age: 73%. Precision production: At age 62: 43%; Before full retirement age: 67%. Operator: At age 62: 49%; Before full retirement age: 77%. Women: Managerial: At age 62: 31%; Before full retirement age: 54%. Professional: At age 62: 25%; Before full retirement age: 49%. Sales: At age 62: 46%; Before full retirement age: 74%. Clerical/Administrative: At age 62: 30%; Before full retirement age: 59%. Services: At age 62: 50%; Before full retirement age: 73%. Health services: At age 62: 27%[A]; Before full retirement age: 68%. Farming: At age 62: N/A; Before full retirement age: N/A. Mechanics: At age 62: N/A; Before full retirement age: N/A. Construction: At age 62: N/A; Before full retirement age: N/A. Precision production: At age 62: N/A; Before full retirement age: N/A. Operator: At age 62: 49%; Before full retirement age: 71%. Source: GAO analysis of Health and Retirement Study data. Note: Before full retirement age includes respondents who claimed from age 62 through 1 month prior to their full retirement age. As denoted by N/A, certain categories did not have sufficient sample size to provide estimates. The estimates do not control for other factors and have relative standard errors of less than 15 percent unless otherwise noted. [A] Relative standard errors are between 15 to 30 percent. [End of figure] In addition to the type of job held, employment status at age 60 through 62 and the length of work history affect whether an individual will claim early benefits. As shown in figure 4, those who are employed full-time at age 60 through 62 are about 30 percent less likely to claim early, after controlling for other factors. Correspondingly, a greater share of those who reported being retired, unemployed, or otherwise out of the labor force claim early-- particularly for those who claim at age 62. As we have previously reported, many unemployed older workers struggle to find new jobs and may face unique reemployment challenges, such as employer reluctance to hire older workers or out-of-date skills. And, long-term unemployment can motivate older workers to claim early Social Security benefits.[Footnote 44] Along with employment status, a worker's length of work history affects whether or not they will claim early. Those who worked at least 35 years by age 60 through 62 were about 38 percent more likely to claim early when compared to those who spent less time in the workforce, after controlling for other factors. Figure 4: Several Work-Related Factors Affect the Likelihood of Claiming Benefits Early after Controlling for Other Factors: [Refer to PDF for image: horizontal bar graph] Worked at least 35 years by age 60-62: More likely: 38%[A]. Blue collar job at age 60-62: More likely: 55%[A]. Less than a college degree: More likely: 23%[B]. Employed full-time at age 60-62: Less likely: 30%[A]. Managerial or professional job at age 60-62: Less likely: 32%[A]. Source: GAO analysis of Health and Retirement Study data. Note: These estimates are based on our multivariate logistic regression analysis described in appendix I. Reference group includes those who did not respond to the survey question. [A] Indicates that this variable is statistically significant at p-value less than 0.01. [B] Indicates that this variable is statistically significant at p-value less than 0.10. [End of figure] Other work-related factors that may pose challenges for continuing to work at older ages are also linked to claiming early. Specifically: * Education--Early claimers tend to have less education. The share of those with less than a college degree claiming early is relatively high compared to those with at least a college education. However, after controlling for other factors, this effect was only marginally significant (i.e., statistically significant at the 90 percent level). Those with less than a college degree were 23 percent more likely to claim early, after controlling for other factors, as shown in figure 4. * 2008 financial crisis--Those who separated from their job in 2009, when the unemployment rate peaked, were much more likely to claim early when compared to those who separated from their jobs in other years, after controlling for other factors. Other recent research has also shown an increase in early claiming following the financial crisis which then declined in the following years.[Footnote 45] Other demographic and financial factors also influence the decision to claim early: In addition to these work-related factors, whether an individual claims early varies by marital status and other demographic and financial factors. In particular, men and women who are widowed are about 3.1 and 4.6 times more likely to claim early, respectively, after controlling for other factors. In contrast, the share of women who are separated, divorced, or partnered that claim early is relatively low; however, those categories were not statistically significant predictors of claiming after controlling for other factors.[Footnote 46] Figure 5 summarizes how the early claiming rate varies by marital status and race for men and women. Figure 5: Share of Those Claiming Social Security Benefits Early by Various Demographic Characteristics: [Refer to PDF for image: 2 horizontal bar graphs] Percentage claiming Social Security benefits: Men: Married: At age 62: 43%; Before full retirement age: 66%. Widowed: At age 62: 59%; Before full retirement age: 89%. Divorced/Separated: At age 62: 45%; Before full retirement age: 62%. Partnered: At age 62: 48%[A]; Before full retirement age: 64%. Never Married: At age 62: 43%[A]; Before full retirement age: 64%. White: At age 62: 44%; Before full retirement age: 66%. Black: At age 62: 42%; Before full retirement age: 58%. Other: At age 62: 41%; Before full retirement age: 62%. Women: Married: At age 62: 45%; Before full retirement age: 64%. Widowed: At age 62: 45%; Before full retirement age: 61%. Divorced/Separated: At age 62: 34%; Before full retirement age: 56%. Partnered: At age 62: 33%[A]; Before full retirement age: 50%. Never Married: At age 62: 48%; Before full retirement age: 65%. White: At age 62: 44%; Before full retirement age: 65%. Black: At age 62: 39%; Before full retirement age: 57%. Other: At age 62: 35%[A]; Before full retirement age: 53%. Source: GAO analysis of Health and Retirement Study data. Note: The share of those who are black or of another race claiming early is significantly lower compared to those who are white. It is important to note, though, that our sample excludes those who are receiving Social Security Disability Insurance, enrollment in which is considerably higher for those who are black or of another race than for whites. The estimates we present have relative standard errors of less than 15 percent unless otherwise noted. [A] Relative standard errors are between 15 to 30 percent. [End of figure] Moreover, early claiming varies by total household wealth (see figure 6). Those in the middle wealth quartiles--greater than or equal to $83,638 to less than $585,052 of total household wealth--are more likely to claim early, after controlling for other factors, compared to those in the lowest and highest wealth quartiles.[Footnote 47] And, we find the same pattern for total household income at age 60 through 62--those in the lowest and highest income quartiles are less likely to claim early compared to those in the middle income quartiles. Figure 6: Share of Those Claiming Social Security Benefits Early by Total Household Wealth Quartile: [Refer to PDF for image: 2 horizontal bar graphs] Percentage claiming Social Security benefits: Men: Lowest quartile: At age 62: 44%; Before full retirement age: 62%. 2nd quartile: At age 62: 49%; Before full retirement age: 73%. 3rd quartile: At age 62: 44%; Before full retirement age: 70%. Highest quartile: At age 62: 41%; Before full retirement age: 64%. Women: Lowest quartile: At age 62: 40%; Before full retirement age: 61%. 2nd quartile: At age 62: 42%; Before full retirement age: 67%. 3rd quartile: At age 62: 46%; Before full retirement age: 68%. Highest quartile: At age 62: 43%; Before full retirement age: 61%. Source: GAO analysis of Health and Retirement Study data. Note: The estimates we present have relative standard errors of less than 15 percent unless otherwise noted. [End of figure] Lower Expectations of Living to Age 75 Are Also Linked to Claiming Early: Expectations of one's own life expectancy also affect whether an individual claims benefits prior to their full retirement age. In particular, if an individual perceives they have a relatively lower chance of living to reach average life expectancy, they may believe it is advantageous to claim early so that they can receive benefits for more years before they die.[Footnote 48] Indeed, an earlier study found that those who claim Social Security benefits early, particularly at age 62, show a higher death rate than those who delay. [Footnote 49] To gauge an individual's expectations about their life expectancy, the HRS asks respondents to report the probability they will live to age 75. After controlling for other factors, we found that those with the highest expectations of living to age 75 are significantly more likely to delay claiming benefits when compared to those with lower expectations.[Footnote 50] However, many of those who claim early still expect a 50 percent or greater chance of living to age 75. Moreover, the extent to which those with the poorest expectations of living to age 75 claim early varies considerably based on their total household wealth. As shown in figure 7, the share of those with the poorest expectations of living to age 75 claiming early is substantially higher for those with total household wealth of about $83,638 or more. Figure 7: Share of Those Claiming Early by Expected Probability of Living to Age 75 and Total Household Wealth Quartile: [Refer to PDF for image: vertical bar graph] Percentage claiming prior to full retirement age: 1st quartile (less than $83,638): Less than 30% expectation of living to age 75: 58%; 30% or higher expectation of living to age 75: 61%. 2nd quartile (more than or equal to $83,638 and less than $236,794): Less than 30% expectation of living to age 75: 75%; 30% or higher expectation of living to age 75: 68%. 3rd quartile (more than or equal to $236,794 and less than $585,052): Less than 30% expectation of living to age 75: 78%; 30% or higher expectation of living to age 75: 68%. 4th quartile (more than or equal to $585,052): Less than 30% expectation of living to age 75: 72%[A]; 30% or higher expectation of living to age 75: 63%. Source: GAO analysis of Health and Retirement Study data. Note: The estimates we present have relative standard errors of less than 15 percent unless otherwise noted. [A] Relative standard errors are between 15 to 30 percent. [End of figure] Differences in life expectancy by factors associated with early and delayed claiming: Certain characteristics that we find make an individual more likely to delay claiming, such as having low income and wealth at age 60 through 62, are also linked to lower life expectancy. In addition, studies examining mortality rates and life expectancy indicate that having less education is linked to lower life expectancy. For certain characteristics, the differences in life expectancy are substantial, even at older ages. Specifically: * Income and wealth--Those in the lowest income and wealth quartiles are more likely to delay claiming compared to those with moderate income and wealth. While the timing of claiming differs for those in the lowest and more moderate income and wealth quartiles, overall, recent research finds a substantial difference in life expectancy by socio-economic status. According to a 2007 study by SSA looking at male Social Security-covered workers in birth cohorts spanning 1912- 1941, the life expectancy at age 60 for those in the bottom half of the lifetime earnings distribution is 5.8 years less than for those in the top half.[Footnote 51] * Race--Although blacks are more likely to delay claiming compared to whites, recent research shows that gaps in life expectancy by race persist.[Footnote 52] Further, this research shows that a considerable amount of the gap in life expectancy by race is associated with differences in education and socio-economic status. * Education--Those with less education are more likely to claim early and have a shorter life expectancy. Recent research has shown that the gap in life expectancy by educational attainment has widened.[Footnote 53] This widening gap is due to there being little change in life expectancy among those who are less educated at the same time there have been considerable improvements in life expectancy among those who are more educated. According to one study using 2008 census data, there is more than a 4-year difference in life expectancy at age 60 for white men with at least some college compared to white men with only a high school education.[Footnote 54] The disparity in life expectancy at age 60 by education also exists for white women and for men and women who are black or Hispanic. While many may be claiming early benefits due to work-related reasons or lower expectations of their life expectancy, others who claim early do not face these challenges. As noted earlier, a large share of people claim benefits prior to their full retirement age, but they are not a uniform group. Among early claimers, about 29 percent thought they had at least a 50 percent chance of living to age 75, and, at age 60 through 62, did not have a job that was physically demanding most or all of the time, and were working full or part-time. However, if we also exclude those who report their job was physically demanding some of the time, the share of early claimers in this group drops to about 15 percent. Those Who Delay Claiming Social Security Have Somewhat Higher Income in Retirement, Even among Those with Similar Income at the Start of Retirement: After claiming benefits, households that delay claiming Social Security retirement benefits until their full retirement age have higher median income and wealth in retirement than those that claim prior to full retirement age. Social Security income also makes up a smaller percentage of household income for delayed claimers than early claimers, although this percentage rises on average for both claiming groups as they age in retirement. Even when we compare households with similar total income soon after claiming, those that delay claiming generally have higher income and wealth later in retirement, and a lower percentage of income from Social Security, than early claimers. Median Income and Wealth Are Mostly Higher for Those Who Delay Claiming Social Security Benefits: Median income and wealth for households that delay claiming Social Security benefits until at least their full retirement age are higher than for those that claim early, according to our analysis of the HRS. More specifically, from 1992 to 2010, households with a respondent who delayed claiming benefits until at least their full retirement age had a median income soon after claiming that was 45 percent higher than households who claimed before full retirement age ($71,907 vs. $49,612 annually, in 2010 dollars; see figure 8).[Footnote 55] Median total household wealth was 25 percent higher for households in which a respondent delayed claiming until at least their full retirement age, compared to those that claimed early ($332,215 vs. $266,660). Moreover, median nonhousing wealth for households that delayed was 63 percent higher than for those that claimed early. Figure 8: Median Household Income and Wealth Statistics, by Claiming Age, after Claiming Benefits (in 2010 Dollars), 1992 to 2010: [Refer to PDF for image: horizontal bar graph] 2010 dollars: Total annual income[A]: Claiming before full retirement age (n=5,609): $49,600; Claiming at full retirement age or later (n=1,281): $71,900. Total wealth: Claiming before full retirement age (n=5,609): $266,700; Claiming at full retirement age or later (n=1,281): $332,200. Non-housing wealth[A]: Claiming before full retirement age (n=5,609): $24,900; Claiming at full retirement age or later (n=1,281): $40,500. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Measures income and wealth in the first HRS wave in which the respondent reports receiving Social Security retirement benefits. [A] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] Households that delay claiming receive a lesser share of their income from Social Security benefits than households that claim early, and have higher median earnings. Despite the higher annual benefit, Social Security benefits accounted for a smaller share of total income for households who delayed claiming until their full retirement age--21.8 percent compared with 28.3 percent of total income for those who claimed early (see figure 9). Earnings from employment appear to account for much of the difference in income across claiming groups-- those who chose to delay claiming until their full retirement age had median earnings from employment of $20,482, as compared to just $3,231 for early claimers. At the same time, however, a higher percentage of those who claimed benefits before their full retirement age also reported receiving pension income than those who delayed.[Footnote 56] Figure 9: Household Income Statistics, by Social Security Claiming Age and Income Source, After Claiming Benefits, 1992 to 2010 (Annual, in 2010 Dollars): [Refer to PDF for image: illustrated table] Median total annual income[A]: Claiming before full retirement age: $49,612; Claiming at full retirement age or later: $71,907. Median annual Social Security benefits[A]: Claiming before full retirement age: $13,956; Claiming at full retirement age or later: $15,989. Median percentage of household income from Social Security[A]: Claiming before full retirement age: 28.3%; Claiming at full retirement age or later: 21.8%. Median annual earnings from employment[A]: Claiming before full retirement age: $3,231; Claiming at full retirement age or later: $20,482. Percentage of households with pension income[A]: Claiming before full retirement age: 49.6%; Claiming at full retirement age or later: 38.2%. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Measures income and wealth in the first HRS wave in which the respondent reports receiving Social Security retirement benefits. Pension income includes not only payment from defined benefit plans but also distributions from defined contribution plans. [A] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] When we take income and wealth profiles of households later in retirement--when the individual turns 72 years old--we find that the financial advantages for those who delay claiming benefits largely hold. Median income measured at this age remains higher for claimers who wait until their full retirement age ($51,200) than for early claimers ($38,400) (see figure 10), but the 33 percent difference in income between the claiming groups is narrower than when measured after claiming.[Footnote 57] Early claimers continue to have a higher median percentage of income coming from Social Security than delayed claimers (see figure 11), although this percentage rose for both claiming groups compared to the measurement after claiming. As was the case earlier in retirement, a greater percentage of households in the early claiming group had some pension income when compared to those who delayed claiming--56.0 percent versus 52.8 percent. At age 72, median earnings from employment from employment are zero for each group. Figure 10: Median Household Income and Wealth Statistics, by Social Security Claiming Age, at Age 72, 1992 to 2010 (in 2010 Dollars): [Refer to PDF for image: horizontal bar graph] 2010 dollars: Total annual income[A]: Claiming before full retirement age (n=2,493): $38,400; Claiming at full retirement age or later (n=802): $51,200. Total wealth: Claiming before full retirement age (n=2,493): $275,900; Claiming at full retirement age or later (n=802): $347,900. Non-housing wealth[A]: Claiming before full retirement age (n=2,493): $29,800; Claiming at full retirement age or later (n=802): $38,700. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Measures income and wealth in the HRS wave in which the respondent reaches age 72. [A] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] Figure 11: Household Income Statistics, by Social Security Claiming Age and Income Source, at Age 72, 1992 to 2010 (Annual, in 2010 Dollars, Except as Noted): [Refer to PDF for image: illustrated table] Median total annual income[A]: Claiming before full retirement age: $38,374; Claiming at full retirement age or later: $51,208. Median annual Social Security benefits[A]: Claiming before full retirement age: $18,598; Claiming at full retirement age or later: $20,929. Median percentage of household income from Social Security[A]: Claiming before full retirement age: 50.9%; Claiming at full retirement age or later: 40.1%. Median annual earnings from employment[A]: Claiming before full retirement age: $0; Claiming at full retirement age or later: $0. Percentage of households with pension income[A]: Claiming before full retirement age: 56.0%; Claiming at full retirement age or later: 52.8%. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Measures income and wealth in the HRS wave in which the respondent reported turning 72. Pension income includes not only payment from defined benefit plans but also distributions from defined contribution plans. [A] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] Compared to Early Claimers with Similar Incomes, Households That Delay Claiming Show Higher Income Later in Retirement: Even among households that had similar total income after claiming, overall those households that delay claiming appear to improve their financial position in retirement relative to those that claim earlier.[Footnote 58] While median income and wealth measure how different claiming groups fare financially at different points of their retirement, these financial differences may have existed in part prior to claiming. Here, we compare income and wealth for households that have similar incomes soon after claiming Social Security, but that claim Social Security at different ages--both early and later in retirement--to determine if claiming at different ages might affect households that enter retirement with similar incomes. Households near median income: For households with total income near the median after claiming Social Security ($51,939 per year, in 2010 dollars), we see a slight relative improvement for delayed claimers overall when compared to early claimers (see figure 12).[Footnote 59] At age 72, average income for the group that delayed claiming was 10 percent higher than for those claiming benefits early.[Footnote 60] Average total wealth and nonhousing wealth for delayed claimers were higher at age 72 than for early claimers, whereas these measures were higher for early claimers soon after claiming. Social Security income, which accounted for a similar percentage of total income at the start of retirement for each claiming group, accounts for a somewhat higher percentage of income for early claimers at age 72. Regarding other sources of income, delayed claimers show higher average earnings from employment in both periods. Early claimers have higher average pension income soon after claiming, but delayed claimers report higher average pension income at age 72, which may in part be due to required minimum distributions from certain retirement plans starting at age 70 ½. Figure 12: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72, 1992 to 2010 (in 2010 Dollars)--Households near Median Income: [Refer to PDF for image: 2 horizontal bar graphs and 2 illustrated tables] After claiming[B]: Total annual income: Claiming before full retirement age (n=1,226): $51,200; Claiming at full retirement age or later (n=185): $51,700. Total wealth: Claiming before full retirement age (n=1,226): $441,300; Claiming at full retirement age or later (n=185): $405,100. Non-housing wealth: Claiming before full retirement age (n=1,226): $114,700; Claiming at full retirement age or later (n=185): $86,900. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 31.7% ($16,044); Claiming at full retirement age or later: 33.3% ($16,973). Households with income from pension (avg. $ amount): Claiming before full retirement age: 60.0%[A] ($11,013[A]); Claiming at full retirement age or later: 40.7% ($5,685). Annual earnings from employment: Claiming before full retirement age: $14,419; Claiming at full retirement age or later: $20,339[A]. At age 72[C]: Total annual income: Claiming before full retirement age (n=535): $50,400; Claiming at full retirement age or later (n=115): $55,400; Total wealth: Claiming before full retirement age (n=535): $424,000; Claiming at full retirement age or later (n=115): $496,100. Non-housing wealth: Claiming before full retirement age (n=535): $132,500; Claiming at full retirement age or later (n=115): $188,800. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 50.2% ($19,665); Claiming at full retirement age or later: 47.4% ($20,565). Households with income from pension (avg. $ amount): Claiming before full retirement age: 62.9% ($10,825); Claiming at full retirement age or later: 60.5% ($13,893). Annual earnings from employment: Claiming before full retirement age: $4,704; Claiming at full retirement age or later: $11,146[A]. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Figure shows averages of households with total income within 20 percent of the median household income for the overall sample ($51,939 in 2010 dollars). [A] Measures income and wealth in the first HRS wave in which the respondent reports receiving Social Security income. [B] Measures income and wealth in the first HRS wave in which the respondent's age is at least 72. [C] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] Households near 10th-and 90th-percentile income: Away from median income levels and closer to the extremes of the income distribution, we see similar results: in many ways, the financial profile of those who delay claiming improves relative to early claimers from the time of claiming to age 72. Average household income for those near 10th-percentile income ($16,629 per year) after claiming rises at age 72 to $28,253 for delayed claimers versus $22,483 for early claimers. Differences in average total and nonhousing wealth between the two claiming groups also diverge over this period. Despite Social Security benefits accounting for a higher percentage of household income after claiming for those who delayed-- 74.7 percent versus 65.8 percent for households that claimed prior their full retirement age--at age 72 early claimers received 74.8 percent of income from benefits, on average, compared to 59.4 percent for delayed claimers. However, early claimers have higher average income from pensions and are more likely to have pension income than delayed claimers. Comparing households near the 90th-percentile income level after claiming ($148,800 per year) shows a more mixed comparison of early vs. later claimers financial outcomes. Average total income at age 72 is over 30 percent higher for those who delay claiming benefits. However, average wealth is higher both early and later in retirement for the early claimer group. (For more details about 10th- and 90th-percentile income households, please see figures in appendix II.) Nearly A Quarter of Early Claimers Lacked Access to Government or Employer-Sponsored Health Coverage and Thus May Benefit From New Options, Mostly at Reduced Costs: Of the 4.2 million early claimers younger than age 65, we estimated that almost 1 million lacked government or employer-sponsored health insurance coverage prior to the recent implementation of new PPACA provisions.[Footnote 61] Of these, 14 percent may be newly eligible for Medicaid in 26 states expanding eligibility. The remaining 86 percent of early claimers without government or employer-sponsored health insurance could choose to purchase coverage through the exchanges, and most could be eligible for federal tax credits to reduce their premiums. We also estimated that 29 percent of early claimers without government or employer-sponsored coverage may not be eligible for these options, including 10 percent with incomes below the poverty level who live in states that did not expand Medicaid and have incomes too low for the federal premium tax credits. (See figure 13 for potential early claimer eligibility for these health coverage options.) The number of early claimers that actually enroll in these new health coverage options could be influenced by individual and employer decisions made in response to the availability of these new options. Figure 13: Early Claimers without Government or Employer-Sponsored Coverage Who May Be Eligible for Expanded Medicaid or Federal Tax Credits for Coverage Purchased through the Exchanges: [Refer to PDF for image: pie-chart with 2 sub-charts] Employer-sponsored or government coverage: 77% (3.3 million); Uninsured or directly purchased coverage: 23% (1.0 million). Of the 1 million: Potentially eligible for Medicaid expansion: 14% (134,630); Could purchase coverage on exchange (not Medicaid eligible): 86% (838,469); Of the 86% of potential exchange users: Eligible for tax credits: 58% (560,832); Income too high for tax credits: 19% (183,330); Income too low for tax credits and too high for Medicaid: 10% (94,307). Source: GAO analysis of U.S. Census Bureau data. Notes: Data are from the 2009-2011 American Community Survey (ACS) 3- year estimates. The estimates we present have relative standard errors of less than 15 percent, and may not add to 100 percent due to rounding. [End of figure] Almost 1 Million Early Claimers Lacked Government or Employer- Sponsored Health Insurance; Most of Whom Were Uninsured: Of the 4.2 million early claimers younger than age 65, we found that 23 percent were not covered by government or employer-sponsored health insurance, according to 2009-2011 ACS data. Of these almost 1 million (973,099) early claimers, more than half had no source of health insurance coverage, and the rest indicated as their only source of coverage that they purchased insurance directly from an issuer offering an individual health plan.[Footnote 62] We found that the remaining 77 percent (3.3 million) of early claimers were covered by government or employer-sponsored health insurance. (See figure 14.) Figure 14: Estimated Early Claimer Health Insurance Coverage in the United States, 2009 through 2011: [Refer to PDF for image: pie-chart] Government (Medicare, Medicaid, military health care, or VA health care): 37% (1.6 million); Employer or union only: 40% (1.7 million); Total: 77% (3.3 million). Insurance purchased directly: 10% (0.4 million); Uninsured: 13% (0.5 million); Total: 23% (1.0 million). Source: GAO analysis of U.S. Census Bureau data. Notes: Data are from the 2009-2011 American Community Survey (ACS) 3- year estimates. The estimates we present have relative standard errors of less than 15 percent, and may not add to total due to rounding. [End of figure] We estimated that early claimers had similar rates of government or employer-sponsored health insurance as those 62 through 64-year-olds who had not claimed Social Security, although the types of coverage varied. Compared to the 77 percent of early claimers with government or employer-sponsored coverage, we estimated that 80 percent of other individuals age 62 through 64 had such coverage. However, a higher percentage of early claimers received their insurance through a government-sponsored program. We estimated that 37 percent of early claimers had coverage under a government-sponsored program, relative to 17 percent for those who had not yet claimed. In contrast, those who had not yet claimed were more likely to have employer-sponsored coverage as their only source of insurance. We estimated that 40 percent of early claimers had employer-sponsored coverage, compared to 64 percent of other individuals this age. Fourteen Percent of Early Claimers without Government or Employer- Sponsored Coverage May Be Newly Eligible for Expanded Medicaid: Of the 973,099 early claimers who were not previously covered by a government-sponsored program or employer-sponsored plan, we estimated that 14 percent (134,630) may be newly eligible for Medicaid in 2014 based on their incomes and applicable federal poverty levels in 2011. These early claimers live in the 26 states that sought to expand Medicaid and received CMS approval as of December 2013. Enrollment in expanded Medicaid would generally provide these individuals with access to insurance with no premiums and no or nominal cost-sharing. [Footnote 63] Across these 26 states, the percentage of these early claimers eligible for Medicaid expansion ranges from 17 percent in Delaware and Massachusetts to 60 percent in the District of Columbia. Across all 26 states, the number of newly eligible early claimers is equal to about 50 percent of the number of early claimers that were already covered by Medicaid. In one state, Nevada, we estimated that there could be as many early claimers eligible for expanded Medicaid as there were early claimers already enrolled in Medicaid. (See appendix II for estimates of potential early claimer eligibility for Medicaid expansion, by state.) In all states, the actual number of early claimers covered by expanded Medicaid in 2014 will depend on actual enrollment. The Remaining 86 Percent of Early Claimers without Government or Employer-Sponsored Coverage May Benefit from Certain PPACA Provisions Designed to Improve Access to the Individual Market in 2014: Many early claimers who were not covered by government or employer- sponsored coverage could benefit from certain PPACA provisions that are designed to improve the availability, affordability, and standardization of coverage offered through the individual market. We estimated that 86 percent (838,469 of 973,099) of these early claimers may not be eligible for expanded Medicaid. These individuals could benefit from certain PPACA provisions that took effect in 2014 and are intended to provide additional protections to individuals obtaining coverage through the individual market--both within and outside of the exchanges--including limits on the extent of variation in their premium rates. Early claimers may choose to purchase coverage through the exchanges in 2014, and most of them may be eligible for premium tax credits to lower the cost of coverage. Plans available through the exchanges: Early claimers will have access to individual health plans through the health insurance exchanges. For all individuals, including early claimers, the number of plans available through the exchange will vary depending on the state and rating area in which they live.[Footnote 64] For example, according to HHS data for 36 states,[Footnote 65] an individual living anywhere in New Hampshire could select among 13 plans offered by one issuer, while an individual living in Florida's most populous rating area could select among 141 plans offered by nine issuers. The number of plans available to an individual also depends on where in a state that individual lives. In Florida, for example, many of the rating areas on the eastern coast of the state had more than 100 available plans, while most of the rating areas in the north- central part of the state each had a total of 27 available plans. The types of plans available to individuals in each exchange also vary by state, and not all plan types are available in all states. According to the HHS data, all individuals living in these 36 states would have access to at least one bronze, silver, and gold plan, regardless of the rating area in which they live. However, some individuals may not have access to the higher premium, lower cost- sharing platinum plans. In 9 states, the exchanges did not offer a platinum level plan in any rating area, and within 10 other states platinum-level plans were offered in some rating areas, but not others. Premiums for plans available through the exchanges: The premiums available to early claimers for individual plans offered through the exchanges vary by state. Nationally, the average premium for a 62-year-old nonsmoker to purchase the second-lowest-cost silver plan available in the rating area we examined is $583 per month, before any tax credits are applied.[Footnote 66] The state with the lowest such premium was Minnesota (about $347 per month), while Connecticut had the highest such premium (almost $946 per month). [Footnote 67] (See figure 15.) While we examined premiums for the most populous county in each state, premiums do vary by rating area within states. For example, in Florida, the monthly premium for the second- lowest-cost silver plan was $605 in its most populous county (Miami- Dade), but varied from about $523 for a county in the western part of the state to about $793 for a county neighboring Miami-Dade. Figure 15: Average Monthly Premiums Paid by a 62-Year-Old Nonsmoker for the Second-Lowest-Cost Silver Plan in Each State: [Refer to PDF for image: illustrated U.S. map] Average monthly premium: $300-500: Arizona; Hawaii; Illinois; Iowa; Kansas; Kentucky; Maryland; Massachusetts; Minnesota; New Mexico; New York; Oklahoma; Oregon; Tennessee; Utah; Vermont. Average monthly premium: $500-500: Alabama; Arkansas; California; Colorado; Delaware; District of Columbia; Florida; Georgia; Idaho; Louisiana; Maine; Michigan; Missouri; Montana; Nebraska; Nevada; New Hampshire; North Carolina; North Dakota; Ohio; Pennsylvania; Rhode Island; South Carolina; South Dakota; Texas; Virginia; Washington; West Virginia. Average monthly premium: $700-900: Alaska; Indiana; New Jersey; Wisconsin; Wyoming. Average monthly premium: more than $900: Connecticut; Mississippi. Source: GAO analysis of Department of Health and Human Services and Henry J. Kaiser Family Foundation data. Note: The premiums presented are what a 62-year-old nonsmoker living in the state's most populous county would pay for the second-lowest-cost silver plan in 2014. The premiums presented represent the amount before any premium tax credits are applied. [End of figure] Early claimers eligible for premium tax credits and cost-sharing subsidies: We estimated, based on 2011 incomes and poverty levels, that 58 percent (560,832) of the 973,099 of early claimers without government or employer-sponsored coverage could meet income requirements for tax credits to lower the cost of premiums for plans purchased through the exchanges. Specifically, these individuals had incomes between 100 and 400 percent FPL and did not qualify for Medicaid or other types of coverage.[Footnote 68] Therefore, if they purchase a second-lowest- cost silver plan through an exchange in their state, they may be eligible to receive a federal tax credit to reduce their premiums to between 2 and 9.5 percent of their income.[Footnote 69] Early claimers living in states with exchange plans that have more expensive premiums will qualify for higher premium tax credits. While the percentage of income that must be paid by individuals towards their premiums remains consistent across states, the amount of the tax credit for each individual will depend on the amount of the premiums. For example, a 62-year-old nonsmoker earning 150 percent FPL would pay just over $57 per month for a second-lowest -cost silver plan purchased in either Minnesota or Connecticut after application of the tax credit. In Minnesota, the federal tax credit of just over $289 would pay for the remainder of the $347 premium; in Connecticut, the tax credit of just over $888 would pay for the remainder of the $946 premium.[Footnote 70] In some states, premiums are low enough that early claimers with incomes near 400 percent FPL may not be eligible for a tax credit.[Footnote 71] (See figure 16 for the monthly premiums that a 62-year-old nonsmoker could pay in the states with the lowest-, median-, and highest-cost plans.) Figure 16: Monthly Premium Paid by a 62-Year-Old Nonsmoker after Applicable Premium Tax Credits for the Second-Lowest-Cost Silver Plan, by Income Level: [Refer to PDF for image: horizontal bar graph] Average monthly premium: State with lowest premium: No tax credit: Paid by tax credit: 0; Paid by consumer: $347. 400% of poverty: Paid by tax credit: 0; Paid by consumer: $347. 150% of poverty: Paid by tax credit: $289; Paid by consumer: $57. State with average premium: No tax credit: Paid by tax credit: 0; Paid by consumer: $580. 400% of poverty: Paid by tax credit: $217; Paid by consumer: $364. 150% of poverty: Paid by tax credit: $523; Paid by consumer: $57. State with highest premium: No tax credit: Paid by tax credit: 0; Paid by consumer: $946. 400% of poverty: Paid by tax credit: $582; Paid by consumer: $364. 150% of poverty: Paid by tax credit: $888; Paid by consumer: $57. Source: GAO analysis of Department of Health and Human Services and Henry J. Kaiser Family Foundation data. Note: These premium tax credit calculations are based on the 2013 federal poverty level (FPL) for an individual living in the 48 contiguous states and the District of Columbia. The FPL for individual living in Alaska or Hawaii is higher. Some totals do not add due to rounding. [End of figure] The value of the tax credits is also likely to be greater for early claimers than younger individuals. Older and younger individuals will be eligible to pay the same income-adjusted premiums, but the unadjusted premiums for older individuals are about 3 times higher. [Footnote 72] For example, on average, the premium for a 21-year-old nonsmoker to purchase a second-lowest-cost silver plan would be $214 per month, while a 62-year-old nonsmoker would pay $583 per month. At these premiums, a 21-year-old nonsmoker earning between 300 and 400 percent FPL would not qualify for a premium tax credit, while a 62- year-old nonsmoker would be eligible for a tax credit of $219. We also estimated that two-thirds (363,796 of 560,823) of the early claimers potentially eligible for premium tax credits may be eligible for federal cost-sharing subsidies for plans purchased through the exchanges. Specifically, these subsidies would lower the cost of out of the pocket expenses, such as copayments and deductibles that individuals would pay when they receive care. (See table 2 for an example of the cost-sharing subsidies for an individual with an income at 150 percent FPL living in California.) Table 2: Example of Benefit Differences under the Cost-Sharing Subsidies: Selected benefit categories: Individual medical deductible; Standard silver plan: $2,000; Silver plan with cost-sharing subsidies for 150% FPL: None. Selected benefit categories: Primary care visit copayment; Standard silver plan: $45; Silver plan with cost-sharing subsidies for 150% FPL: $3. Selected benefit categories: Specialty care visit copayment; Standard silver plan: $65; Silver plan with cost-sharing subsidies for 150% FPL: $5. Selected benefit categories: Maximum out of pocket for one; Standard silver plan: $6,350; Silver plan with cost-sharing subsidies for 150% FPL: $2,250. Source: GAO review of plan information available through the California health insurance exchange. Note: The federal government makes payments to issuers to cover the reductions in cost-sharing. This is intended to allow qualifying individuals to reduce their cost-sharing, without purchasing a more expensive gold or platinum plan. The plans presented in this table are an example of plans offered through the exchange of one state. The specific benefits of other plans will vary, within the requirements set by PPACA. [End of table] Almost One-Third of Early Claimers without Government or Employer- Sponsored Coverage May Not Be Eligible for Medicaid or Premium Tax Credits: We estimated that almost one-third of early claimers without government or employer-sponsored coverage, including some with incomes below the poverty level, may not be eligible for either expanded Medicaid or the premium tax credits for exchange plans. Specifically, we estimated that 10 percent (94,307 of 973,099) of early claimers without government or employer-sponsored coverage had incomes below 100 percent FPL ($11,490 in most states in 2013), based on 2011 incomes and poverty levels. Specifically, these early claimers lived in one of the 25 states not approved to expand Medicaid in 2014 and did not reach or exceed the 100 percent FPL threshold to qualify for premium tax credits.[Footnote 73] In addition to that 10 percent of early claimers, we estimated that another 19 percent (183,330) of early claimers without government or employer-sponsored coverage had incomes too high to qualify for premium tax credits, over 400 percent FPL. Without tax credits, the premiums for coverage available through the exchanges could represent a relatively large percentage of income for these populations. For example, the average premium for a 62-year-old nonsmoker living in one of the 25 non-expansion states to purchase the second-lowest-cost silver plan was $628 per month. For an individual with income at 99 percent FPL, this premium would be equivalent to almost 66 percent of their income. This individual could also choose to purchase a lower-cost bronze plan. For a 62-year-old nonsmoker earning 99 percent FPL in one of the 25 non-expansion states, the average lowest-cost bronze plan would cost $469 per month, or about 49 percent of their income. Similarly, early claimers with incomes just over 400 percent FPL would not benefit from the tax credits. For example, for a 62-year-old nonsmoker earning 401 percent FPL, the average premium of the second- lowest-cost silver plan would be equivalent to 15.1 percent of their income. That percentage would decrease as the person's income increased. For example, at 700 percent FPL, that person would pay an average of 8.7 percent of their income to purchase this type of plan. (See figure 17 for the average percent of income paid to premiums for a 62-year-old nonsmoker after applicable premium tax credits.) Figure 17: Average Percent of Income Paid to Premiums for a 62-Year- Old Nonsmoker after Applicable Premium Tax Credits for the Second- Lowest-Cost Silver Plan, by Income Level: [Refer to PDF for image: line graph] Percentage of poverty level/Annual income (2013 dollars): 100%; $11,490: Percentage of income spent on premiums: 2%. Percentage of poverty level/Annual income (2013 dollars): 150%; $17,235: Percentage of income spent on premiums: 4%. Percentage of poverty level/Annual income (2013 dollars): 250%; $28,725: Percentage of income spent on premiums: 8.1%. Percentage of poverty level/Annual income (2013 dollars): 400%; $45,960: Percentage of income spent on premiums: 9.5%. Percentage of poverty level/Annual income (2013 dollars): 401%: Percentage of income spent on premiums: 15.1%. Percentage of poverty level/Annual income (2013 dollars): 500%; $57,450: Percentage of income spent on premiums: 12.2%. Percentage of poverty level/Annual income (2013 dollars): 600%; $68,940: Percentage of income spent on premiums: 10.1%. Percentage of poverty level/Annual income (2013 dollars): 700%; $80,430: Percentage of income spent on premiums: 8.7%. Source: GAO analysis of Department of Health and Human Services and Henry J. Kaiser Family Foundation data. Note: PPACA includes a federal premium tax credit for eligible individuals whose incomes are between 100 and 400 percent of the federal poverty level (FPL). Within this range, individuals would pay no more than a fixed percentage of their income in premiums. Above 400 percent FPL, individuals would pay the full premium. These premium tax credit calculations are based on the 2013 FPL in the 48 contiguous states and the District of Columbia. The FPL for individuals living in Alaska or Hawaii is higher. This figure shows the percentage of income that 62-year-olds at various income levels would pay for a plan with a premium of $581 per month, the average premium for a 62-year-old nonsmoker to purchase the second-lowest-cost silver plan for the rating areas we examined in the 48 contiguous states and the District of Columbia. [End of figure] Actual Enrollment of Early Claimers in These New Health Coverage Options Could Be Influenced by Other Changes in the Health Insurance Market: The number of early claimers that actually enroll in these new health coverage options in 2014 could be influenced by individual and employer decisions made in response to the availability of these new options. Our estimates of early claimer eligibility for new coverage options are based on coverage from 2009 through 2011; we did not measure actual enrollment. It is too early to assess if the availability of new coverage options through the individual market will affect these trends in 2014 and beyond. For example, prior to PPACA, some individuals, such as those with preexisting conditions, may have deferred retiring and claiming Social Security at age 62 in order to maintain their employer-sponsored insurance.[Footnote 74] With the availability of new options through the individual market, more individuals between the ages of 62 and 64 could decide to retire and claim early.[Footnote 75] In addition, while employers had already been reducing coverage for active or retired employees, some may decide to terminate coverage once the PPACA provisions are fully implemented. In 2012, we reported that between 9 and 20 percent of employers were considering dropping coverage for retirees in response to PPACA.[Footnote 76] Similarly, in a 2011 survey, Mercer LLC found that while relatively few employers expected to terminate plans after the exchanges became operational in 2014, a much larger proportion considered them to be a viable alternative for their retirees.[Footnote 77] In its 2012 survey, Mercer LLC found that only about half of large employers currently sponsoring retiree health plans expected they would be offering such plans in 5 years.[Footnote 78] Further, the individual health insurance market outside of the exchanges is changing in ways that could affect participation in the exchanges. We estimated that about 45 percent (442,093 of 973,099) of early claimers without government or employer-sponsored coverage were covered by health insurance that they purchased directly through the individual market, according to 2009-2011 ACS data. For our analysis, we considered these individuals to be eligible to purchase coverage through the exchanges. However, some of these individuals may also be able to maintain the plan they had prior to 2014, rather than purchasing new coverage through the exchanges. But, in some cases, insurers canceled individual market plans prior to 2014 because the plans did not meet PPACA requirements that took effect in 2014. Individuals seeking new coverage in 2014 could decide to purchase coverage through the exchange or outside the exchanges. Concluding Observations: Even as more Americans delay claiming Social Security benefits until reaching their full retirement age, a considerable number of people may still find it difficult to work until older ages and will therefore claim benefits early. Our findings highlight some of the challenges of devising policy related to Social Security claiming. For many workers, delaying claiming Social Security carries important financial benefits and ensures a higher minimum level of retirement income regardless of how long retirement lasts; further, raising the full retirement age appears to have influenced many workers to delay claiming. The overall trend toward delayed claiming may continue as those approaching retirement adjust to the need to spread their retirement savings over more years due to increases in average life expectancy. However, workers who face physically demanding working conditions or who have lost their jobs late in their careers may lack the flexibility to delay claiming. In this way, early access to Social Security acts as an important safety net for older workers with poor job prospects. Our results also highlight the key role Social Security plays in providing retirement security, with Social Security benefits usually making up an increasing share of income as households age in retirement across all income levels. As pensions that pay lifetime benefits become less available, retirees may face an even greater need to maximize Social Security benefits in order to provide an income floor if retirement lasts longer than expected. However, while our findings demonstrate the financial advantages of delaying claiming until the full retirement age, this does not necessarily mean that everyone would be better off delaying. The optimal claiming decision for each individual depends on many factors, such as expected mortality, employment opportunities, and health. For those individuals who find it difficult to work at older ages and claim Social Security before becoming eligible for Medicare, certain PPACA provisions could provide them with additional, affordable coverage options that improve retirement security. At the same time, these new health coverage options add a layer of complexity to an already difficult financial decision. Moreover, for those who do not qualify for federal tax credits for exchange plans, especially those living below poverty in states that did not expand Medicaid, coverage purchased through the individual market could still represent a significant portion of their income. PPACA could also have effects beyond those examined in this report if worker and employer decisions about retirement change in response to the availability of these new options. While our findings indicate that PPACA should improve the availability of affordable coverage options, it is too early to know the ultimate effect of these new provisions on individual decisions of when to retire. Agency Comments: We provided a draft of this report to the Department of Health and Human Services, the Department of Labor, and the Social Security Administration for review and comment. The Department of Health and Human Services stated that it had no comments. The Department of Labor and the Social Security Administration provided technical comments which we incorporated, as appropriate. As agreed with your office, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies of this report to the Secretary of Health and Human Services, the Secretary of Labor, the Commissioner of Social Security, and other interested parties. In addition, the report will be available at no charge on the GAO website at [hyperlink, http://www.gao.gov]. If you or your staff have any questions about this report, please contact John E. Dicken at (202) 512-7114 or dickenj@gao.gov or Charles Jeszeck at (202) 512-7215 or jeszeckc@gao.gov. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. GAO staff who made contributions to this report are listed in appendix V. Sincerely yours, Signed by: John E. Dicken: Director, Health Care: Signed by: Charles A. Jeszeck: Director, Education, Workforce, and Income Security: [End of section] Appendix I: Objectives, Scope, and Methods: To analyze the circumstances faced by those who claim early Social Security retirement benefits, we examined the following questions: (1) What demographic and occupational characteristics are associated with claiming Social Security benefits before the full retirement age? (2) How does the composition and distribution of retirement income for those who claim Social Security retirement benefits early compare to those who delay? (3) To what extent did early claimers lack health insurance prior to 2014, and how have their coverage options changed since the implementation of the Patient Protection and Affordable Care Act (PPACA). This appendix provides a detailed account of the information and methods we used to answer these questions. Section 1 describes the key information sources and empirical methods we used to answer questions 1 and 2, as well as the results of supplementary analyses. Sections 2 and 3 provide the same information for question 3. For each objective, we reviewed relevant federal laws and regulations. Section 1: Examining the Characteristics, Income, and Wealth of Early Claimers: To answer our questions, we obtained information from the Health and Retirement Study (HRS), reviewed relevant literature, and conducted interviews with experts and agency officials. Health and Retirement Study: To answer questions 1 and 2, we analyzed data collected through the HRS, a nationally representative survey primarily sponsored by the National Institute on Aging and conducted by the Institute for Social Research at the University of Michigan. This longitudinal survey collects information on individuals over age 50 and contains detailed data on their education, marital status, work history, health, assets, and income. Respondents are surveyed every 2 years. Additional cohorts are added over the course of the survey, which was first conducted in 1992, to maintain the representation of the older population. The data in our analysis include respondents from various survey waves from 2000 through 2010. The sample selection criteria varies for objectives 1 and 2 based on the nature of the information needed to answer these questions as described in the corresponding methods sections that follow. One of the main advantages of the HRS is that the same households are interviewed at different points in time, allowing us to capture information before and after life events, such as claiming Social Security benefits, occur. Further, RAND, a research organization, cleans and processes the HRS data to create a user-friendly longitudinal dataset that has consistent and intuitive naming conventions and model-based imputations for missing wealth and income data. However, there are limitations in using these data for our analysis. In particular, the data are self-reported. As such, they are subject to the respondent's perceptions and possible errors in identifying specific figures, such as their assets and income, as well as the date on which they first received Social Security benefits. Data Reliability: For the HRS, we conducted a data reliability assessment of selected variables by conducting electronic data tests for completeness and accuracy, reviewing documentation on the dataset, or interviewing knowledgeable officials about how the data are collected and their appropriate uses. When we learned that particular fields were not sufficiently reliable, we did not use them in our analysis. For example, we chose not to use data where there were a substantial number of missing values. We also checked to see that values for variables were internally consistent (such as spot checking that values for formulas matched the values of its component variables) and that results were not affected unduly by outlier values that might suggest miscoded values. For the purposes of our analysis, we found the variables we ultimately reported on to be sufficiently reliable. Literature Review and Interviews: To gain an understanding of the circumstances faced by those who claim early Social Security benefits, we conducted a literature review and interviewed experts and agency officials. We conducted a search of various publications using the Proquest database. In addition, we coordinated with the Congressional Research Service (CRS) and Congressional Budget Office (CBO) to identify relevant studies, and checked with agency officials as to whether they would recommend any additional materials. Officials from the Social Security Administration (SSA) recommended and supplied dozens of relevant studies, which we included in our review. Finally, during interviews with experts, we asked for recommendations for other noteworthy studies. Our review primarily focused on studies from the last five years (2008-2013), but we also included earlier studies that were very relevant. We reviewed article abstracts and identified those which were most relevant to our research objectives. Methods for Examining Demographic and Occupational Characteristics of Early and Delayed Claimers: To determine which factors are associated with claiming Social Security benefits prior to the full retirement age, we reviewed relevant literature and interviewed experts and agency officials to discuss their research methodologies and findings. Based on this review, we analyzed data from the HRS from 2000 to 2010. Using HRS data, we examined key characteristics of early and delayed claimers, such as demographics, health status, and work history. Sample selection: We selected respondents observed during their early claiming window-- between age 62 to 1 month prior to their full retirement age at the time of the survey. To focus our analysis on the claiming of Social Security retirement benefits, we applied several exclusion criteria in constructing our sample. Specifically, we excluded the following observations from our sample: * Individuals who received Social Security Disability Insurance benefits; * Individuals who reported claiming benefits prior to age 62; * Individuals who report a marital status of widowed prior to claiming benefits; * Individuals where neither they nor their spouse (if married) reported working at least 10 years; * Individuals where they and their spouse (if married) worked in public administration for their longest tenure job;[Footnote 79] and: * Individuals who died between age 62 and full retirement age. The remaining individuals in our sample are most likely to be receiving retired worker or spousal benefits, which are the focus of this analysis. To eliminate multiple observations of the same individual across different waves, we selected the observation for the earliest wave in which the individual reported claiming benefits. For those who have not yet claimed by their full retirement age, we selected the observation to retain based on a random sample after stratifying by age group (i.e., age 62, 63, and 64 or older) to mimic the age distribution of those who claim early. Data censoring: The issue of censored data, whereby we do not observe all respondents throughout their entire early retirement window, emerges in the later waves, particularly in the 2008 and 2010 waves. Without adjusting our sample, the incidence of claiming early is biased downward because someone who has not yet claimed, but may eventually claim early (e.g., in 2011), would have the early claiming indicator set to "0." On the other hand, dropping all observations where we do not observe someone claim and we do not observe them throughout their entire early claiming window could provide an upward bias in the incidence of early claiming. To address this issue, we compared the incidence of early claiming in HRS data from the 2006, 2008, and 2010 waves to SSA data for comparable birth cohorts. By comparing the HRS and SSA data, we observed the impact of the downward bias from the censoring issues described previously. We determined that applying the following criteria mitigated the effect of censoring and resulted in an incidence of early claiming that is comparable to SSA's data: (1) in the 2008 wave, drop observations where the respondent is age 62 and has not yet claimed Social Security, and (2) in the 2010 wave, drop observations where the respondent is age 62 to 63 and has not yet claimed Social Security. After applying our exclusion criteria, the resulting sample contained 5,779 observations. Use of complex design and sample weight: As described by HRS survey design document,[Footnote 80] the original HRS core sample design was a multistage area probability sample of households based on the Survey Research Center's 84 strata National Sample frame. To allow independent analysis of key subgroups, the core sample was augmented by three supplements. These supplements were an oversample of African Americans, an oversample of Hispanics, and an oversample of Floridians. In addition, in order to improve coverage of the oldest old households with extremely frail respondents, a second sampling frame was employed: the Health Care Financing Administration Enrollment Data Base. To correct for the differential sampling rates of these subgroups and for differential non-response, sampling weights are provided for each wave of data for use in making inferences to the population. The HRS sample is selected under a multistage area probability sample design. The sample includes four distinct selection stages. An overview of these selection stages is given here. The primary stage of sampling involves probability proportionate to size selection of U.S. Metropolitan Statistical Areas and non-Metropolitan Statistical Areas counties. This stage is followed by a second stage sampling of area segments (SSU) within sampled primary stage units. The third stage of sample selection is preceded by a complete listing (enumeration) of all housing units that are physically located within the bounds of the selected SSU. The third sampling stage is a systematic selection of housing units from the housing unit listings for the sample SSUs. The fourth and final stage in the multistage design is the selection of an age-eligible person within a sample housing unit.[Footnote 81] In order to have an accurate representation of a sample unit, we took account of multistage clustered sample design and sample weights in both descriptive statistics and multivariate analysis. Table 3: Sample Statistics: Factor: Overall; Sample count: 5,779; Percent of sample: 100%; Proportion of sample claiming at age 62: 42.67%; Proportion of sample claiming before full retirement age: 64.73%. Factor: Gender: Male; Sample count: 2,661; Percent of sample: 46%; Proportion of sample claiming at age 62: 43.25%; Proportion of sample claiming before full retirement age: 65.35%. Factor: Gender: Female; Sample count: 3,118; Percent of sample: 54%; Proportion of sample claiming at age 62: 42.17%; Proportion of sample claiming before full retirement age: 64.21%. Factor: Education: Less than college degree; Sample count: 4,492; Percent of sample: 78%; Proportion of sample claiming at age 62: 45.55%; Proportion of sample claiming before full retirement age: 67.07%. Factor: Education: College degree or higher; Sample count: 1,270; Percent of sample: 22%; Proportion of sample claiming at age 62: 32.36%; Proportion of sample claiming before full retirement age: 56.38%. Factor: Marital status: Married; Sample count: 4,621; Percent of sample: 80%; Proportion of sample claiming at age 62: 43.48%; Proportion of sample claiming before full retirement age: 64.79%. Factor: Marital status: Separated or divorced; Sample count: 722; Percent of sample: 12%; Proportion of sample claiming at age 62: 36.43%; Proportion of sample claiming before full retirement age: 58.03%. Factor: Marital status: Widowed; Sample count: 276; Percent of sample: 5%; Proportion of sample claiming at age 62: 45.65%; Proportion of sample claiming before full retirement age: 82.61%. Factor: Marital status: Partnered; Sample count: 195; Percent of sample: 3%; Proportion of sample claiming at age 62: 38.46%; Proportion of sample claiming before full retirement age: 54.36%. Factor: Marital status: Never married; Sample count: 157; Percent of sample: 3%; Proportion of sample claiming at age 62: 42.04%; Proportion of sample claiming before full retirement age: 62.42%. Factor: Ethnic group: White; Sample count: 4,764; Percent of sample: 82%; Proportion of sample claiming at age 62: 43.8%1; Proportion of sample claiming before full retirement age: 66.31%. Factor: Ethnic group: Black; Sample count: 788; Percent of sample: 14%; Proportion of sample claiming at age 62: 37.94%; Proportion of sample claiming before full retirement age: 57.49%. Factor: Ethnic group: Hispanic; Sample count: 505; Percent of sample: 9%; Proportion of sample claiming at age 62: 39.41%; Proportion of sample claiming before full retirement age: 60.79%. Factor: Ethnic group: Other race; Sample count: 227; Percent of sample: 4%; Proportion of sample claiming at age 62: 35.24%; Proportion of sample claiming before full retirement age: 56.83%. Factor: Total wealth: Q1: Less than $83,638; Sample count: 1,140; Percent of sample: 20%; Proportion of sample claiming at age 62: 38.68%; Proportion of sample claiming before full retirement age: 60.53%. Factor: Total wealth: Q2: $83,638 - less than $236,794; Sample count: 1,142; Percent of sample: 20%; Proportion of sample claiming at age 62: 43.78%; Proportion of sample claiming before full retirement age: 69.61%. Factor: Total wealth: Q3: $236,794 - less than $585,052; Sample count: 1,142; Percent of sample: 20%; Proportion of sample claiming at age 62: 46.32%; Proportion of sample claiming before full retirement age: 69.26%. Factor: Total wealth: Q4: $585,052+; Sample count: 1,142; Percent of sample: 20%; Proportion of sample claiming at age 62: 42.56%; Proportion of sample claiming before full retirement age: 63.92%. Factor: Total wealth: Missing; Sample count: 1,213; Percent of sample: 21%; Proportion of sample claiming at age 62: *; Proportion of sample claiming before full retirement age: *. Factor: Occupation category: Professional, managerial; Sample count: 1,014; Percent of sample: 18%; Proportion of sample claiming at age 62: 26.33%; Proportion of sample claiming before full retirement age: 53.35%. Factor: Occupation category: Blue collar; Sample count: 729; Percent of sample: 13%; Proportion of sample claiming at age 62: 47.46%; Proportion of sample claiming before full retirement age: 75.45%. Factor: Occupation category: Others; Sample count: 4,036; Percent of sample: 70%; Proportion of sample claiming at age 62: 45.9%1; Proportion of sample claiming before full retirement age: 65.66%. Factor: Expectations of living to age 75: Less than 30%; Sample count: 428; Percent of sample: 7%; Proportion of sample claiming at age 62: 45.33%; Proportion of sample claiming before full retirement age: 65.65%. Factor: Expectations of living to age 75: 30% - less than 50%; Sample count: 155; Percent of sample: 3%; Proportion of sample claiming at age 62: 50.97%; Proportion of sample claiming before full retirement age: 66.45%. Factor: Expectations of living to age 75: 50% - less than 90%; Sample count: 2,789; Percent of sample: 48%; Proportion of sample claiming at age 62: 48.48%; Proportion of sample claiming before full retirement age: 69.74%. Factor: Expectations of living to age 75: 90%+; Sample count: 1,426; Percent of sample: 25%; Proportion of sample claiming at age 62: 42.71%; Proportion of sample claiming before full retirement age: 63.04%. Factor: Expectations of living to age 75: Missing; Sample count: 981; Percent of sample: 17%; Proportion of sample claiming at age 62: *; Proportion of sample claiming before full retirement age: *. Source: GAO analysis of HRS data. Note: Due to respondents with missing values for some categories, the total sample count may not add up to 5,779 and the percent may not add up to 100. Also, overlapping categories in marital status and ethnic group may result in over 100 percent. [End of table] Data analysis: We analyzed descriptive statistics and developed a standard statistical model to determine the association of these factors with early retirement. In the course of our analysis, we considered two categories of early claimers--(1) those who claimed in the year in which they turned age 62 and (2) those who claimed before their full retirement age, which was increased from 65 to 66 during the time period encompassed by our sample. We determined that using the second definition (i.e., claiming before the full retirement age) provided more inclusive results from statistical modeling. For comparison, we provide the descriptive statistics for both definitions of early claimers in the findings section.[Footnote 82] Respondent level weights were applied in estimating descriptive statistics and regression analysis to adjust for the HRS oversampling certain sub- populations (e.g., African American, Hispanics, and residents of Florida). Some variables of interest, such as financial and health status variables, may be affected by the age at which we observe the respondent. For example, if we compare the health of someone who claimed at age 62 with the health of someone who claimed at age 65, the latter respondent might have worse health simply because they are older, rather than that difference in health causing one individual to claim early and the other to delay. To mitigate possible age effects, we captured information for most variables when respondents were approximately the same age, just before or at the time they first are eligible to claim Social Security benefits (i.e., the wave in which they are between ages 60 to 62). Regression analysis: To develop our final statistical model, we conducted a series of bivariate analyses to estimate the associations between the different factors (or independent variables) and the dependent variable (i.e., a dummy variable indicating whether or not the respondent claimed early). We used separate indicators for claiming at age 62 or claiming at any time prior to their full retirement age. Based on these analyses, we determined which variables were significantly associated with early claiming. In some cases, we identified independent variables that appeared to be statistically significant predictors of an individual's claiming status, but omitted them out of concern that they may have been more an "effect" of the individual's claiming status than a "cause" of it. For example, the current income of an individual may be driven by whether or not they claimed Social Security benefits and retired from their job, rather than their current income affecting their decision to claim early or delay. Where we identified such concerns, we either created a variable to capture that factor prior to claiming (i.e., the wave in which the individual was age 60 to 62) or eliminated the variable. We also examined correlation matrices of selected variables to determine where there were high correlations between two variables. Where variables were highly correlated, we developed an indicator that combined the two variables or chose one variable over the other. Based on these results, we tried various specifications for our model to determine which combination of variables provided the most meaningful results and the greatest explanatory power. To estimate our model, we used logistic regression--an appropriate technique when the dependent variable is binary, or has two categories such as claiming early or delaying. Logistic regression also allows for the coefficients to be converted into odds ratios, which are described below. Table 4 shows the results of the logistic regression that demonstrate the relationship between the independent variables with the probability of claiming benefits prior to the full retirement age. The first column presents the variables that were included in the model. The third column presents odds ratios that are estimated for each variable in the model.[Footnote 83] Odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to claim early. For example, an odds ratio of 1.55 for those who worked in a blue collar occupation between the ages of 60 to 62 would mean that they are 1.55 times more likely to claim benefits prior to their full retirement age. Odds ratios that are significantly lower than 1.00 indicate that individuals with that characteristic are less likely to claim early. Table 4: Factors Associated with Claiming Social Security Benefits Prior to the Full Retirement Age: Independent variables: Female Widowed; Proportion of early claimers: 89.1; Odds Ratio from multivariate model: 4.60; Statistical significance: [C]. Independent variables: Male Widowed; Proportion of early claimers: 81.5; Odds Ratio from multivariate model: 3.09; Statistical significance: [C]. Independent variables: White; Proportion of early claimers: 66.0; Odds Ratio from multivariate model: 1.40; Statistical significance: [C]. Independent variables: Veteran; Proportion of early claimers: 68.9; Odds Ratio from multivariate model: 1.35; Statistical significance: [C]. Independent variables: Self-assessed health status fair/poor, at age between 60 and 62; Proportion of early claimers: 60.4; Odds Ratio from multivariate model: 0.65; Statistical significance: [C]. Independent variables: Covered by a health insurance, at age between 60 and 62; Proportion of early claimers: 62.9; Odds Ratio from multivariate model: 0.69; Statistical significance: [C]. Independent variables: Covered by a government health plan, at age between 60 and 62; Proportion of early claimers: 48.0; Odds Ratio from multivariate model: 0.38; Statistical significance: [C]. Independent variables: Probability of life expectancy exceed 75 years old > 90%, at age between 60 and 62; Proportion of early claimers: 61.2; Odds Ratio from multivariate model: 0.83; Statistical significance: [B]. Independent variables: Total asset in the 2nd and 3rd quartile, at age between 60 and 62; Proportion of early claimers: 69.5; Odds Ratio from multivariate model: 1.29; Statistical significance: [C]. Independent variables: Total household income in the 2nd and 3rd quartile, at age between 60 and 62; Proportion of early claimers: 69.5; Odds Ratio from multivariate model: 1.29; Statistical significance: [C]. Independent variables: High school or less education; Proportion of early claimers: 67.3; Odds Ratio from multivariate model: 1.23; Statistical significance: [A]. Independent variables: Full-time labor force status, at age between 60 and 62; Proportion of early claimers: 62.1; Odds Ratio from multivariate model: 0.70; Statistical significance: [C]. Independent variables: Managerial or professional occupation, at age between 60 and 62; Proportion of early claimers: 52.5; Odds Ratio from multivariate model: 0.68; Statistical significance: [C]. Independent variables: Blue colored occupation, at age 60 and 62; Proportion of early claimers: 75.9; Odds Ratio from multivariate model: 1.55; Statistical significance: [C]. Independent variables: Worked 35 years or longer; Proportion of early claimers: 69.2; Odds Ratio from multivariate model: 1.38; Statistical significance: [C]. Independent variables: Last year worked = 2000; Proportion of early claimers: 76.1; Odds Ratio from multivariate model: 2.11; Statistical significance: [C]. Independent variables: Last year worked = 2001; Proportion of early claimers: 79.2; Odds Ratio from multivariate model: 2.55; Statistical significance: [C]. Independent variables: Last year worked = 2002; Proportion of early claimers: 73.9; Odds Ratio from multivariate model: 1.91; Statistical significance: [C]. Independent variables: Last year worked = 2003; Proportion of early claimers: 80.0; Odds Ratio from multivariate model: 2.97; Statistical significance: [C]. Independent variables: Last year worked = 2004; Proportion of early claimers: 79.9; Odds Ratio from multivariate model: 2.74; Statistical significance: [C]. Independent variables: Last year worked = 2005; Proportion of early claimers: 80.0; Odds Ratio from multivariate model: 2.76; Statistical significance: [C]. Independent variables: Last year worked = 2006; Proportion of early claimers: 84.0; Odds Ratio from multivariate model: 3.80; Statistical significance: [C]. Independent variables: Last year worked = 2007; Proportion of early claimers: 84.1; Odds Ratio from multivariate model: 3.63; Statistical significance: [C]. Independent variables: Last year worked = 2008; Proportion of early claimers: 86.9; Odds Ratio from multivariate model: 4.42; Statistical significance: [C]. Independent variables: Last year worked = 2009; Proportion of early claimers: 88.8; Odds Ratio from multivariate model: 6.30; Statistical significance: [C]. Independent variables: Last year worked = 2010; Proportion of early claimers: 85.5; Odds Ratio from multivariate model: 4.01; Statistical significance: [C]. Source: GAO analysis of HRS data. Notes: Proportion estimates are weighted by sample weights and odds ratio are estimated from multivariate model reflecting complex survey design and sample weights. Reference group includes respondents who did not respond.Since we applied several exclusion criteria from the original sample design and data, standard error estimates for the subpopulation tend to be slightly lower than the domain estimate. However, the magnitude of such difference is so small and did not materially impact to the level of significance to any factor in the model. [A] Indicates that the variable is statistically significant at the 90 percent level (p-value less than 0.1). [B] Indicates that the variable is statistically significant at the 95 percent level (p-value less than 0.05). [C] Indicates that the variable is statistically significant at the 99 percent level (p-value less than 0.01). [End of table] Model goodness-of-fit: Because of the lack of goodness-of-fit for a logistic regression model fitted using survey sample data with complex design, we examined this by first fitting the design-based model (i.e., one that takes the survey design structure into account), then estimating the corresponding probabilities, and subsequently using independently and identically distributed-based tests and applying any findings to the design based model.[Footnote 84] The statistical properties of this procedure and an alternative goodness-of-fit test for logistic regression when modeling data collected using sample survey data have been studied previously.[Footnote 85] Unlike ordinary goodness-of-fit tests, this alternative test takes into account the sampling weights and design.[Footnote 86] After fitting the logistic regression model taking the survey sampling design into account, the F-adjusted mean residual goodness-of-fit test was applied and suggested no evidence of lack of fit. Methods for Examining the Composition and Distribution of Retirement Income for Early and Delayed Claimers: To determine the composition and distribution of retirement income for early and delayed claimers, we compared income, wealth, and income sources of early and delayed Social Security claimers. We selected our sample of eligible claimers using the same selection criteria as objective 1 (such as removing all Social Security Disability Insurance recipients).[Footnote 87] We then assign each eligible claimer to the same "early" (claimed benefits prior to the respondent's full retirement age) and "delayed" claiming groups. We take comparisons of the two claiming group's income and wealth at two different points in time: (1) in the first HRS wave after the respondent has claimed Social Security benefits, and (2) in the first HRS wave in which the respondent is 72 years old. This allows us to compare income and wealth of the two claiming groups both (for those still in the HRS sample at age 72) early and later in retirement. To identify the wave in which a respondent first has claimed Social Security, we determine the first wave in which the household reports income from Social Security benefits after the respondent has reported claiming Social Security; we then measure all income and wealth statistics in that wave for the respondent's household. (Note that this compares early and delayed claimers at different ages, because by definition these groups begin to receive retirement benefits at different ages.) For the age-72 wave, we identify the wave in which the respondent turns 72, and measure income and wealth statistics for the respondent's household in that wave. Finally, we do two separate types of comparisons of the different claiming group's income and wealth. 1. Medians--For each group of claimers, we measure median total income, total wealth, and total nonhousing wealth. We also measure median levels of different income sources, for early and delayed claimers in the wave following claiming and in the age-72 wave. Income sources we track are Social Security benefits, earnings, pensions, capital income, and other income.[Footnote 88] The unit of measure is the household; for variables the HRS keeps at individual levels only (such as the components of income), we add respondent and spouse values together.[Footnote 89] 2. Average income and wealth of households with similar income early in retirement--This analysis selects only households with similar total income in the wave after claiming, and compares average (not median) income and wealth both in the early and age-72 waves for the claiming groups. More specifically, we identify median, 10th- percentile, and 90th-percentile total household income for the combined group of Social Security claimers (both early and delayed claimers together), measured during the early post-claiming HRS wave. We then select households with total income within 20 percent of these different points on the income distribution.[Footnote 90] We take average income and wealth for each claiming group, in both the early and later retirement waves. We use averages instead of medians for this analysis because averages give a better measure of typical income sources and wealth for a household with similar levels of income. While averages can be affected by outliers, we think that choosing a banded range of households near the same income levels limit this effect. Further, this analysis shows us how households who choose to claim benefits at different ages, but who enter retirement with similar income might, differ financially later in retirement. For all analyses we convert all dollar values to 2010 dollars (because 2010 is the most recent HRS wave we use) to control for inflation. We do not, however, adjust for cohort effects (real income and benefits of retirees from later waves should be higher than those from earlier waves). Section 2: Early Claimer Health Insurance Coverage: To estimate the number of early claimers potentially eligible for expanded Medicaid and premium tax credits and cost-sharing subsidies for individual plans offered through the health exchanges in each state beginning in 2014, we examined data from the American Community Survey (ACS) of the U.S. Census Bureau. ACS collects information about age, sex, race, family and relationships, income and benefits, health insurance, education, veteran status, disabilities, and geographic location, among other variables. The ACS initially samples about 3 million addresses each year, resulting in about 2 million final interviews for each year's survey. For the 2011 survey, a sample of 2.3 million households was selected from a sample frame of about 3.5 million. For the 2010 and 2009 surveys, a sample of 2.1 million households was selected from a sample frame of 3.1 million, for each year. In general, ACS estimates are period estimates that describe the average characteristics of population and housing over a period of data collection. The 2009-2011 ACS 3-year estimates are averages over the period from January 1, 2009, to December 31, 2011, and reflect what the average value is over the full period. Because survey data are based on a sample of the population, the numbers in the report are not the same as those that would be obtained if everyone in the population were actually counted. For this reason, the ACS combines data from multiple years to produce reliable estimates for small areas--such as states--and population groups. To produce data for small communities--between 20,000 and 65,000 individuals--3 years of data collection are pooled to produce reliable estimates. Estimates from the ACS are subject to some sampling error. To assess the precision of these estimates, we calculated a relative standard error for each estimate. A relative standard error is calculated by dividing the standard error of the estimate by the estimate itself. For example, if an estimate has a mean of 100 and a standard error of 20, the relative standard error would be 20/100, which would be 20 percent. Estimates with small relative standard errors are considered more reliable than estimates with large relative standard errors. A small relative standard error is a more precise measurement since there is less variance around the mean. Unless otherwise noted, we present estimates that have relative standard errors of less than 15 percent. In addition to sampling errors, nonsampling errors may affect the data; however, the Census Bureau takes steps to minimize these errors. The results of ACS data analysis are based on self-reported information. The Census Bureau applied some logical edits to the data to correct for obvious misunderstanding of the questions. Like other survey-based sources of coverage data--such as the National Health Interview Survey--the estimates from ACS data of Medicaid coverage are lower than numbers found in program counts by the Centers for Medicare & Medicaid Services (CMS), even after logical edits have been applied by the Census Bureau. Reporting errors can be due to several factors, including reluctance to report public coverage or confusion about what type of insurance respondents have. In order to account for a variety of scenarios that might contribute to the differences in these numbers, one group of researchers has developed an additional set of edits--for example, assigning to Medicaid a child for whom no coverage was reported when that child has a sibling enrolled in Medicaid.[Footnote 91] We did not apply these edits because the publicly available data set was sufficient for the purpose of providing estimates of early claimer eligibility. We assessed the reliability of 2009-2011 ACS data by reviewing documentation and discussing the database with knowledgeable officials. After taking these steps we determined the data were sufficiently reliable for our purposes. Augmented American Community Survey Dataset: In order to capture the appropriate family ties and income for determining program eligibility, we used an augmented version of the ACS data prepared by the University of Minnesota Population Center-- the Integrated Public Use Microdata Series.[Footnote 92] Because coverage options often consider family ties in determining program eligibility, we used the "health insurance unit," a variable included in the Integrated Public Use Microdata Series data and considered the most accurate unit of analysis for estimating health insurance eligibility using survey data. The health insurance unit consists of all individuals residing in a sampled household, whose family ties determine them to be eligible for coverage, whereas the definition of a family or household in the Census's nonaugmented data typically includes all related members of a household, not all of whom would be eligible for coverage.[Footnote 93] Estimating Early Claimer Insurance Coverage: We identified early claimers as those who received Social Security retirement income.[Footnote 94] For these analyses, we excluded individuals older than 64, because they would be eligible for Medicare when they turn 65. We also identified the types of insurance coverage reported by early claimers. ACS asks respondents if they are currently covered by one of several forms of health insurance or health coverage plans. The options include insurance through an employer or union, Medicare, Medicaid, TRICARE or other military health care, Department of Veterans Affairs (VA), purchased directly from an insurance company, and the Indian Health Service. We categorized respondents as being without government or employer-sponsored health insurance coverage if they purchased coverage directly through the individual market, [Footnote 95] had coverage through the Indian Health Service,[Footnote 96] or were uninsured and had no other source of coverage. Estimating Early Claimers' Eligibility for Expanded Medicaid and Tax Credits and Cost Subsidies for Exchange Plans: We estimated the number of early claimers without government or employer-sponsored health insurance coverage who may be eligible for expanded Medicaid and federal tax credits and cost-sharing subsidies in 2014. Under the Patient Protection and Affordable Care Act (PPACA), a uniform measure of income, known as modified adjusted gross income (MAGI), must be used to determine eligibility for these provisions. [Footnote 97] MAGI excludes income received from public welfare assistance, Supplemental Security Income, veterans' benefits, employment compensation, child support, and certain other categories of income. While ACS does not include a measure of MAGI, it does include data on income received from public welfare assistance and Supplemental Security Income. Therefore, we estimated MAGI by subtracting these sources of income from an individual's total income. We were not able to use ACS to identify the other sources of excluded income, such as veterans benefits and child support. Income was adjusted to 2011 dollars across all 3 years of the survey, in order to determine the poverty level status for program eligibility.[Footnote 98] To estimate the total number of early claimers who are likely to be eligible for expanded Medicaid, we used the 2009-2011 data to identify those early claimers without government or employer-sponsored coverage who live in a state that chose to expand Medicaid, and met the income eligibility standard for an adult with no dependent children, generally at or below 133 percent federal poverty level (FPL).[Footnote 99] CMS had approved the following states to expand Medicaid as of December 2013: Arizona, Arkansas, California, Colorado, Connecticut, Delaware, District of Columbia, Hawaii, Illinois, Iowa, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Jersey, New Mexico, New York, North Dakota, Ohio, Oregon, Rhode Island, Vermont, Washington, and West Virginia. We also estimated the number of early claimers potentially eligible for the tax credits and cost-sharing subsidies for coverage purchased through the exchanges beginning in 2014. Specifically, for those early claimers without government or employer-sponsored coverage who did not meet the income eligibility standard for Medicaid expansion, we estimated the number who met the income requirements for the premium tax credits (100 to 400 percent FPL) or cost-sharing subsidies (100 to 250 percent FPL). In this report, we do not conduct projections of eligibility based on future income; instead, we estimate the number of early claimers who would be eligible for programs and benefits beginning in 2014 based on the most recent data on income and coverage currently available. We did not predict the number of people who will participate in these programs and benefits, commonly referred to as the take-up rate. As public programs generally do not have full take-up, it is likely not all who are eligible for Medicaid or the premium tax credit and cost- sharing subsidies will enroll. Program eligibility among our samples was determined by comparing income of the health insurance unit to the FPL. In order to determine whether individuals were eligible for expanded Medicaid or the premium tax credit and cost-sharing subsidies for exchange plans, we defined two factors needed to compare the income to the FPL: (1) the number of people in the family--defined by the health insurance unit in our analysis; and (2) the family income, added for all individuals with the same health insurance unit identification code. We used a similar methodology as part of our recent report examining the eligibility of American Indians and Alaska Natives for expanded Medicaid and the premium tax credit and cost-sharing subsidies for exchange plans.[Footnote 100] Section 3: Early Claimer Health Insurance Exchange Premiums: We also obtained data on premiums for individual health plans offered through the exchanges. For the 36 states with exchanges operated or supported by the federal government, we examined the Department of Health and Human Services' (HHS) Health Insurance Marketplace Premiums for 2014 Databook.[Footnote 101] For those 15 exchanges operated by the states, complete premium data were not available from HHS. Therefore, we used a web-based calculator developed by the Henry J. Kaiser Family Foundation.[Footnote 102] We assessed the reliability of these data through review of related documentation, by selectively comparing these premiums with other sources, and through a discussion with a Henry J. Kaiser Family Foundation official. After taking these steps we determined the data were sufficiently reliable for our purposes. In each state, we used the appropriate tool to identify the premiums for a nonsmoker age 62 and, for comparison, age 21. Within each state, premiums vary across geographic areas, known as rating areas. To assure the premiums we report represent what a large number of individuals in the state would pay, we selected the premiums offered in the most populous county using the most currently available U.S. Census data.[Footnote 103] To use the HHS databook, we identified the rating area in which that county is located; to use the Henry J. Kaiser Family Foundation calculator, we identified a ZIP code associated with that county and rating area. [End of section] Appendix II: Average Household Income and Wealth for Households near 10th and 90th Percentile of Total Household Income: Figure 18: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72 (in 2010 Dollars)-- Households near 10th Percentile Total Household Income: [Refer to PDF for image: 2 horizontal bar graphs and 2 illustrated tables] After claiming[B]: Total annual income: Claiming before full retirement age (n=437): $16,800; Claiming at full retirement age or later (n=57): $17,100. Total wealth: Claiming before full retirement age (n=437): $204,500; Claiming at full retirement age or later (n=57): $236,800. Non-housing wealth: Claiming before full retirement age (n=437): $26,400; Claiming at full retirement age or later (n=57): $52,900[A]. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 65.8% ($11,063); Claiming at full retirement age or later: 74.7%[A] ($12,564[A]). Households with income from pension (avg. $ amount): Claiming before full retirement age: 22.6% ($1,219); Claiming at full retirement age or later: 13.5% ($786). Annual earnings from employment: Claiming before full retirement age: $2,100[A]; Claiming at full retirement age or later: $1,359. At age 72[C]: Total annual income: Claiming before full retirement age (n=187): $22,500; Claiming at full retirement age or later (n=44): $28,300[A]; Total wealth: Claiming before full retirement age (n=187): $181,700; Claiming at full retirement age or later (n=44): $461,400[A]. Non-housing wealth: Claiming before full retirement age (n=187): $27,900; Claiming at full retirement age or later (n=44): $184,400[A]. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 74.8% ($14,425); Claiming at full retirement age or later: 59.4%[A] ($14,078). Households with income from pension (avg. $ amount): Claiming before full retirement age: 37.7% ($4,037); Claiming at full retirement age or later: 26.0% ($2,989). Annual earnings from employment: Claiming before full retirement age: $1,656; Claiming at full retirement age or later: $2,392. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Figure includes averages of households with total income within 20 percent of the 10th-percentile household income for the overall sample ($16,629 in 2010 dollars). [A] Measures income and wealth in the first HRS wave in which the respondent reports receiving Social Security income. [B] Measures income and wealth in the first HRS wave in which the respondent's age is at least 72. [C] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] Figure 19: Average Household Income and Wealth Statistics, by Social Security Claiming Age, after Claiming and at Age 72 (in 2010 Dollars)-- Households near 90th Percentile Total Household Income: [Refer to PDF for image: 2 horizontal bar graphs and 2 illustrated tables] After claiming[B]: Total annual income: Claiming before full retirement age (n=364): $143,400; Claiming at full retirement age or later (n=154): $143,800. Total wealth: Claiming before full retirement age (n=364): $1,177,500; Claiming at full retirement age or later (n=154): $1,012,800. Non-housing wealth: Claiming before full retirement age (n=364): $314,400; Claiming at full retirement age or later (n=154): $446,3. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 11.0% ($15,739); Claiming at full retirement age or later: 13.5%a[A] ($19,270[A]). Households with income from pension (avg. $ amount): Claiming before full retirement age: 55.3% ($21,929[A]); Claiming at full retirement age or later: 44.8% ($12,403). Annual earnings from employment: Claiming before full retirement age: $52,023; Claiming at full retirement age or later: $66,674[A]. At age 72[C]: Total annual income: Claiming before full retirement age (n=147): $89,400; Claiming at full retirement age or later (n=88): $117,400; Total wealth: Claiming before full retirement age (n=147): $1,283,000; Claiming at full retirement age or later (n=88): $1,085,000. Non-housing wealth: Claiming before full retirement age (n=147): $540,000; Claiming at full retirement age or later (n=88): $416,800. Annual income from Social Security (% of households, avg. $ amount): Claiming before full retirement age: 36.2% ($22,612); Claiming at full retirement age or later: 35.1% ($25,251). Households with income from pension (avg. $ amount): Claiming before full retirement age: 67.9% ($23,947); Claiming at full retirement age or later: 51.6% ($19,982). Annual earnings from employment: Claiming before full retirement age: $9,265; Claiming at full retirement age or later: $31,614[A]. Source: GAO analysis of Health and Retirement Study, 1992-2010. Note: Figure includes averages of households with total income within 20 percent of the 90th-percentile household income for the overall sample ($148,800 in 2010 dollars). [A] Measures income and wealth in the first HRS wave in which the respondent reports receiving Social Security income. [B] Measures income and wealth in the first HRS wave in which the respondent's age is at least 72. [C] Indicates that the difference by claiming group in this variable is statistically significant at p-value less than 0.05. [End of figure] [End of section] Appendix III: Estimates of the Number and Percentage of Early Claimers Potentially Eligible for Expanded Medicaid in 2014: Table 5 provides information on the estimated number of early claimers who may be eligible for expanded Medicaid in 2014. Specifically, 26 states were approved to expand Medicaid to cover non-pregnant adults who are not eligible for Medicare and have income up to 133 percent of the federal poverty level (FPL).[Footnote 104] For each of these states, the table presents the number of early claimers with Medicaid coverage according to coverage trends from 2009 through 2011. For those early claimers who were not covered by Medicaid or another form of government or employer-sponsored coverage, we estimated the number who may be eligible, based on their income, for expanded Medicaid in 2014. Table 5: Estimated Early Claimer Eligibility for Medicaid in Those States Approved to Expand Medicaid, by State: State: Arizona; Prior early claimer enrollment in Medicaid[A]: Number: 10,042; Percent of early claimers: 10.4%; Potentially eligible for expanded Medicaid: Number: 6,411; Percent of early claimers without government or employer-sponsored coverage: 26.0%; Percentage increase from prior enrollment: 63.8%. State: Arkansas; Prior early claimer enrollment in Medicaid[A]: Number: 6,764; Percent of early claimers: 12.4%; Potentially eligible for expanded Medicaid: Number: 4,579; Percent of early claimers without government or employer-sponsored coverage: 36.1%; Percentage increase from prior enrollment: 67.7%. State: California; Prior early claimer enrollment in Medicaid[A]: Number: 60,425; Percent of early claimers: 16.6%; Potentially eligible for expanded Medicaid: Number: 32,105; Percent of early claimers without government or employer-sponsored coverage: 34.3%; Percentage increase from prior enrollment: 53.1%. State: Colorado; Prior early claimer enrollment in Medicaid[A]: Number: 6,200; Percent of early claimers: 10.8%; Potentially eligible for expanded Medicaid: Number: 3,320; Percent of early claimers without government or employer-sponsored coverage: 22.1%; Percentage increase from prior enrollment: 53.5%. State: Connecticut; Prior early claimer enrollment in Medicaid[A]: Number: 5,449; Percent of early claimers: 13.3%; Potentially eligible for expanded Medicaid: Number: 2,617; Percent of early claimers without government or employer-sponsored coverage: 33.1%; Percentage increase from prior enrollment: 48.0%. State: Delaware[B]; Prior early claimer enrollment in Medicaid[A]: Number: 1,710; Percent of early claimers: 11.4%; Potentially eligible for expanded Medicaid: Number: 434; Percent of early claimers without government or employer-sponsored coverage: 17.0%; Percentage increase from prior enrollment: 25.4%. State: District of Columbia[B]; Prior early claimer enrollment in Medicaid[A]: Number: 1,494; Percent of early claimers: 38.5%; Potentially eligible for expanded Medicaid: Number: 422; Percent of early claimers without government or employer-sponsored coverage: 59.9%; Percentage increase from prior enrollment: 28.2%. State: Hawaii[B]; Prior early claimer enrollment in Medicaid[A]: Number: 1,815; Percent of early claimers: 10.5%; Potentially eligible for expanded Medicaid: Number: 1,241; Percent of early claimers without government or employer-sponsored coverage: 37.6%; Percentage increase from prior enrollment: 68.4%. State: Illinois; Prior early claimer enrollment in Medicaid[A]: Number: 15,948; Percent of early claimers: 11.0%; Potentially eligible for expanded Medicaid: Number: 10,167; Percent of early claimers without government or employer-sponsored coverage: 30.7%; Percentage increase from prior enrollment: 63.8%. State: Iowa[B]; Prior early claimer enrollment in Medicaid[A]: Number: 5,066; Percent of early claimers: 11.3%; Potentially eligible for expanded Medicaid: Number: 2,048; Percent of early claimers without government or employer-sponsored coverage: 21.0%; Percentage increase from prior enrollment: 40.4%. State: Kentucky; Prior early claimer enrollment in Medicaid[A]: Number: 10,401; Percent of early claimers: 13.0%; Potentially eligible for expanded Medicaid: Number: 5,742; Percent of early claimers without government or employer-sponsored coverage: 34.8%; Percentage increase from prior enrollment: 55.2%. State: Maryland; Prior early claimer enrollment in Medicaid[A]: Number: 5,987; Percent of early claimers: 9.7%; Potentially eligible for expanded Medicaid: Number: 3,795; Percent of early claimers without government or employer-sponsored coverage: 31.6%; Percentage increase from prior enrollment: 63.4%. State: Massachusetts[B]; Prior early claimer enrollment in Medicaid[A]: Number: 17,587; Percent of early claimers: 26.8%; Potentially eligible for expanded Medicaid: Number: 1,504; Percent of early claimers without government or employer-sponsored coverage: 16.5%; Percentage increase from prior enrollment: 8.6%. State: Michigan; Prior early claimer enrollment in Medicaid[A]: Number: 19,112; Percent of early claimers: 10.4%; Potentially eligible for expanded Medicaid: Number: 9,242; Percent of early claimers without government or employer-sponsored coverage: 26.1%; Percentage increase from prior enrollment: 48.4%. State: Minnesota; Prior early claimer enrollment in Medicaid[A]: Number: 8,935; Percent of early claimers: 12.3%; Potentially eligible for expanded Medicaid: Number: 4,611; Percent of early claimers without government or employer-sponsored coverage: 29.8%; Percentage increase from prior enrollment: 51.6%. State: Nevada; Prior early claimer enrollment in Medicaid[A]: Number: 3,139; Percent of early claimers: 8.4%; Potentially eligible for expanded Medicaid: Number: 3,200; Percent of early claimers without government or employer-sponsored coverage: 33.5%; Percentage increase from prior enrollment: 101.9%. State: New Jersey; Prior early claimer enrollment in Medicaid[A]: Number: 10,057; Percent of early claimers: 10.0%; Potentially eligible for expanded Medicaid: Number: 5,773; Percent of early claimers without government or employer-sponsored coverage: 29.7%; Percentage increase from prior enrollment: 57.4%. State: New Mexico; Prior early claimer enrollment in Medicaid[A]: Number: 4,146; Percent of early claimers: 12.6%; Potentially eligible for expanded Medicaid: Number: 2,724; Percent of early claimers without government or employer-sponsored coverage: 31.5%; Percentage increase from prior enrollment: 65.7%. State: New York; Prior early claimer enrollment in Medicaid[A]: Number: 40,229; Percent of early claimers: 16.0; Potentially eligible for expanded Medicaid: Number: 11,219; Percent of early claimers without government or employer-sponsored coverage: 30.9; Percentage increase from prior enrollment: 27.9. State: North Dakota[B]; Prior early claimer enrollment in Medicaid[A]: Number: 670; Percent of early claimers: 7.4%; Potentially eligible for expanded Medicaid: Number: [Empty]; Percent of early claimers without government or employer-sponsored coverage: [Empty]; Percentage increase from prior enrollment: [Empty]. State: Ohio; Prior early claimer enrollment in Medicaid[A]: Number: 17,673; Percent of early claimers: 10.2%; Potentially eligible for expanded Medicaid: Number: 10,154; Percent of early claimers without government or employer-sponsored coverage: 28.0%; Percentage increase from prior enrollment: 57.5%. State: Oregon; Prior early claimer enrollment in Medicaid[A]: Number: 6,348; Percent of early claimers: 9.4%; Potentially eligible for expanded Medicaid: Number: 4,521; Percent of early claimers without government or employer-sponsored coverage: 23.3%; Percentage increase from prior enrollment: 71.2%. State: Rhode Island[B]; Prior early claimer enrollment in Medicaid[A]: Number: 1,610; Percent of early claimers: 12.6%; Potentially eligible for expanded Medicaid: Number: 1,223; Percent of early claimers without government or employer-sponsored coverage: 41.1%; Percentage increase from prior enrollment: 76.0%. State: Vermont[B]; Prior early claimer enrollment in Medicaid[A]: Number: 1,866; Percent of early claimers: 19.2%; Potentially eligible for expanded Medicaid: Number: [Empty]; Percent of early claimers without government or employer-sponsored coverage: [Empty]; Percentage increase from prior enrollment: [Empty]. State: Washington; Prior early claimer enrollment in Medicaid[A]: Number: 9,593; Percent of early claimers: 10.4%; Potentially eligible for expanded Medicaid: Number: 4,791; Percent of early claimers without government or employer-sponsored coverage: 20.2%; Percentage increase from prior enrollment: 49.9%. State: West Virginia; Prior early claimer enrollment in Medicaid[A]: Number: 4,895; Percent of early claimers: 121%; Potentially eligible for expanded Medicaid: Number: 2,158; Percent of early claimers without government or employer-sponsored coverage: 34.4%; Percentage increase from prior enrollment: 44.1%. State: Total; Prior early claimer enrollment in Medicaid[A]: Number: 277,161; Percent of early claimers: 13.0%; Potentially eligible for expanded Medicaid: Number: 134,001; Percent of early claimers without government or employer-sponsored coverage: 29.3%; Percentage increase from prior enrollment: 48.3%. Source: GAO analysis of U.S. Census Bureau data. Notes: Data are from the 2009-2011 American Community Survey (ACS) 3- year estimates. This table presents data for those early claimers from age 62 through 64, too young to qualify for Medicare based on their age. Our estimates are for those states that are approved to expand Medicaid. We did not predict the number of people who will participate in this program, commonly referred to as the take-up rate. Unless otherwise noted, the estimates we present have relative standard errors of less than 15 percent. [A] We refer to prior enrollment as the estimated enrollment based on the responses to the 2009-2011 ACS. [B] Relative standard error of some estimates for this state are between 15 and 30 percent. Estimates with relative standard errors greater than 30 percent are omitted. [End of table] [End of section] Appendix IV: Comments from the Social Security Administration: Social Security: Office of the Commissioner: Social Security Administration: Baltimore, MD 21235-000: March 28, 2014: Mr. Charles A. Jeszeck, Director: Education, Workforce, and Income Security Issues: United States Government Accountability Office: 441 G. Street, NW: Washington, DC 20548: Dear Mr. Jeszeck, Thank you for the opportunity to review the draft report, "Retirement Security: Challenges for Those Claiming Social Security Benefits Early and New Health Coverage Options" (GAO-14-311). We have no comments. If you have any questions, please contact me at (410) 966-9014. Your staff may contact Gary S. Hatcher, Senior Advisor for Records Management and Audit Liaison Staff, at (410) 965-0680. Sincerely, Signed by: Katherine Thornton: Deputy Chief of Staff: Enclosure: [End of section] Appendix V: GAO Contacts and Staff Acknowledgments: GAO Contacts: John Dicken, (202) 512-7114 or dickenj@gao.gov: Charles Jeszeck, (202) 512-7215 or jeszeckc@gao.gov: Staff Acknowledgments: In addition to the contact named above, Michael Collins (Assistant Director), Randy DiRosa, (Assistant Director), Will Hadley (Assistant Director), Mark Glickman, Shannon Grabich, Cynthia Grant, Sharon Hermes, Kirsten Lauber, Ying Long, Sheila McCoy, Susan Offutt, Liam O'Laughlin, Dae Park, Christine San, Hemi Tewarson, Roger Thomas, and Frank Todisco made key contributions to this report. [End of section] Footnotes: [1] Unless it is otherwise clear from context, for purposes of this report, we refer to "early claimers" as those who claim Social Security benefits between the age of 62 and their full retirement age and "delayed claimers" as those who claim Social Security benefits upon or after reaching their full retirement age. [2] The decision to claim Social Security benefits may or may not coincide with retiring from work. Some workers may decide to stop work before becoming eligible for Social Security while others may decide to continue working even after claiming benefits, which may reduce benefits (depending on the amount of earnings) until the full retirement age is reached, but may also lead to a higher future benefit amount. [3] Medicare is the federal health insurance program for seniors, certain individuals with disabilities, and individuals with end-stage renal disease. [4] In 2012, only 24 percent of large employers offered such coverage to new employees, compared with 46 percent in 1993. Mercer, LLC, National Survey of Employer-Sponsored Health Plans: 2012 Survey Report (New York, N.Y.: 2013). [5] Pub. L. No. 111-148, 124 Stat. 119 (2010), as amended by the Health Care and Education Reconciliation Act of 2010, Pub. L. No. 111- 152, 124 Stat. 1029. For purposes of this report, references to PPACA include the amendments made by the Health Care and Education Reconciliation Act of 2010. [6] Medicaid is the joint federal-state health care program that finances health care coverage for certain categories of low-income adults and children. In this report, the term "state" includes the District of Columbia. [7] Because we exclude those who receive Social Security disability benefits, our sample is likely to be healthier on average than the overall U.S. population. [8] In order to capture the appropriate family ties and income for determining program eligibility, we used an augmented version of the ACS data prepared by the University of Minnesota Population Center-- the Integrated Public Use Microdata Series. Steven Ruggles et al., Integrated Public Use Microdata Series: Version 5.0, machine-readable database (Minneapolis: University of Minnesota, 2010). [9] PPACA provides for states to expand Medicaid to cover nonelderly adults who do not qualify for Medicare and who have income up to 133 percent of the federal poverty level (FPL). PPACA also provides for a 5 percent income disregard when calculating income for determining Medicaid eligibility, which effectively increases this income level to 138 percent of the FPL. The FPL is based on federal poverty guidelines issued by the Department of Health and Human Services (HHS) on an annual basis. In this report, we refer to states that opted to expand their programs to cover this population as "expanded Medicaid." A premium tax credit subsidizes the cost of premiums for plans purchased through the exchanges by eligible tax filers and their families. Cost-sharing subsidies reduce the amount eligible individuals or their families would otherwise have to pay when receiving a service (such as deductibles or copayments). To qualify, an individual or family must generally have income between 100 and 400 percent of FPL (for the premium tax credit) or between 100 and 250 percent of FPL (for cost-sharing subsidies) and not qualify for other health care coverage, such as Medicare, Medicaid, or employer- sponsored plans. To examine early claimers' eligibility for expanded Medicaid or subsidies for exchange plans, we developed an estimate of modified adjusted gross income (MAGI), which is a tax-based definition of income. PPACA generally requires the use of MAGI to determine eligibility for these programs. MAGI excludes certain sources of income, such as public welfare assistance and Supplemental Security Income. We excluded income received from both these sources. We were not able to use ACS to identify other sources of excluded income, such as veterans benefits and child support. [10] HHS, Health Insurance Marketplace Premiums for 2014 Databook, downloaded October 30, 2013, [hyperlink, http://aspe.hhs.gov/health/reports/2013/marketplacepremiums/datasheet_ho me.cfm]; Henry J. Kaiser Family Foundation, Subsidy Calculator: Premium Assistance for Coverage in Exchanges, accessed January 16, 2014, http://kff.org/interactive/subsidy-calculator/]. [11] GAO, Private Health Insurance: The Range of Base Premiums for Individuals Age 19 and 64 in the Individual Market by State in January 2013, [hyperlink, http://www.gao.gov/products/GAO-14-263R] (Washington, D.C.: Jan. 31, 2014). [12] For purposes of this report, we describe a select number of PPACA provisions that either expand health insurance options or provide additional protections to individuals seeking health insurance through the individual market. We did not, however, identify all PPACA provisions that may potentially benefit early claimers. [13] Social Security calculates monthly retirement benefits typically by applying a formula to up to 35 years of a worker's highest average indexed monthly earnings, indexed to reflect general wage levels during the years of employment. In addition, a spouse of an eligible worker may be eligible for a benefit based on the worker's earnings, worth up to half of the worker's benefit amount. Spousal benefits are also reduced if claimed prior to the spouse's full retirement age. [14] While SSA defines full retirement age as the age at which an individual is entitled to unreduced benefits, individuals who choose to delay claiming beyond their full retirement age receive delayed retirement credits, until age 70, that increase their benefit amount. For those born in 1943 or later, delayed retirement credits amount to 8 percent per year. [15] The Social Security Amendments of 1983 raised the full retirement age from age 65, starting with retirees born in 1938, gradually up to age 67 for those born in 1960 or later. Pub. L. No. 98-21, § 201, 97 Stat. 65, 107-108. [16] According to SSA, an individual must be at least age 62 "throughout the month" in order to receive a benefit. Coupled with the rule that for benefit entitlement purposes a person attains a specific age the day before their birthday, that means only individuals born on the 1ST or 2ND of the month could claim benefits in the first month they turn age 62. Others born on the 3RD of the month or later could not claim benefits until age 62 and 1 month. [17] Under the current formula, the benefit reduction is 5/9 of 1 percent for each month up to 36 months before the full retirement age and then increased 5/12 of 1 percent for each additional month before the full retirement age. [18] GAO, Retirement Decisions: Federal Policies Offer Mixed Signals about When to Retire, [hyperlink, http://www.gao.gov/products/GAO-07-753] (Washington, D.C.: July 2007). [19] Survivor benefits are based on the deceased worker's benefits. If the deceased claimed benefits prior to his full retirement age, his surviving spouse's benefits would be based on that reduced monthly benefit. If the deceased waited to claim until his full retirement age or later, the surviving spouse could receive the unreduced benefit once he or she reached full retirement age. [20] Psychological factors include feelings of fairness and ownership or loss aversion, whereby forgoing early benefit payments is viewed as a potential loss. See, for example: Suzanne B. Shu and John W. Payne, "Life Expectation Judgments, Fairness, and Loss Aversion in the Psychology of Social Security Claiming Decisions," Working Paper, August 2013 and Melissa A.Z. Knoll, "Behavioral and Psychological Aspects of the Retirement Decision," Social Security Bulletin, Vol. 71, No. 4, 2011. [21] GAO, Social Security Statements: Observations on SSA's Plans for the Social Security Statement, [hyperlink, http://www.gao.gov/products/GAO-11-787T] (Washington, D.C.: Jul. 8, 2011). [22] S. Jay Olshansky, et. al., "Differences in Life Expectancy Due to Race and Educational Differences Are Widening, and Many May Not Catch Up," Health Affairs, 31, No. 8 (2012): 1803-1813. [23] Ellen R. Meara, Seth Richards and David M. Cutler, "The Gap Gets Bigger: Changes in Mortality and Life Expectancy, by Education, 1981- 2000", Health Affairs, 27, No. 2 (2008): 350-360. [24] In addition, experts have highlighted the gains from delay based on the impact on survivor benefits. See, for example, William Meyer and William Reichenstein, "Social Security: When to Start Benefits and How to Minimize Longevity Risk," Journal of Financial Planning. Vol. 23(3), Mar. 2010. John B. Shoven and Sita Nataraj Slavov, "The Decision to Delay Social Security Benefits: Theory and Evidence," NBER Working Paper 17866, Feb. 2012. Owen Haaga and Richard W. Johnson, "Social Security Claiming: Trends and Business Cycle Effects," Center for Retirement Research at Boston College, working paper 2012-5, 2012. [25] Early claimers may have employer-sponsored insurance through retiree coverage sponsored by a previous employer. Alternatively, early claimers may continue to work or opt for coverage under the Consolidated Omnibus Budget Reconciliation Act of 1985 (COBRA). Under COBRA, employers must offer coverage to former employees, such as retirees. [26] Those under age 65 who receive Social Security Disability Insurance generally become eligible for Medicare after a 2-year waiting period, while those diagnosed with end-stage renal disease or amyotrophic lateral sclerosis may become eligible for Medicare without a 2-year waiting period. [27] Active duty military personnel, retirees, and others may be eligible for health care services through the Department of Defense's TRICARE program. Veterans who served active military duty and were discharged or released under conditions other than dishonorable are generally eligible for health care services through the Veterans Health Administration. [28] In 2012, GAO reported that 12 states had a guaranteed issue requirement and 8 states had requirements that restricted the extent to which insurers could adjust premiums based on individual characteristics. See GAO, Private Health Insurance: Estimates of Individuals with Pre-Existing Conditions Range from 36 Million to 122 Million, [hyperlink, http://www.gao.gov/products/GAO-12-439] (Washington, D.C.: Mar. 27, 2012). [29] America's Health Insurance Plans, Individual Health Insurance 2009: A Comprehensive Survey of Premiums, Availability, and Benefits (Washington, D.C.: 2009). In 2011, we reported on this study and other data on application and coverage denials. See GAO, Private Health Insurance: Data on Application and Coverage Denials, [hyperlink, http://www.gao.gov/products/GAO-11-268] (Washington, D.C.: Mar. 16, 2011). [30] A preexisting condition is a health condition that exists before someone applies for or enrolls in a new health insurance policy. See [hyperlink, http://www.gao.gov/products/GAO-12-439]. [31] [hyperlink, http://www.gao.gov/products/GAO-14-263R]. Because insurers were able to impose different rates on individuals for a variety of factors, such as health status, gender, and age, the data represent base premium amounts prior to underwriting, rather than the actual premium amount an individual may have been charged, which could be higher. [32] [hyperlink, http://www.gao.gov/products/GAO-12-439]. [33] Some of these market reforms do not apply to grandfathered plans. A grandfathered health plan refers to an existing plan in which at least one individual has been enrolled since March 23, 2010. To maintain grandfathered status, a plan must avoid certain changes to benefits, cost-sharing, employer contributions, and access to coverage. [34] Issuers offering individual coverage must accept every individual who applies for coverage in that state but may limit this to open or special enrollment periods. In addition, issuers offering individual coverage are prohibited from conditioning eligibility on health status- related factors. 42 U.S.C, §§ 300gg-1, 300gg-4. [35] 42 U.S.C, § 300gg. Plans are also able to vary premiums based on family size and the geographic area in which the enrollee lives. [36] 42 U.S.C. § 18022(a), (d). [37] States administer and operate their Medicaid programs in accordance with state plans that must be approved by the Centers for Medicare & Medicaid Services (CMS). [38] CMS approved the following states to expand Medicaid: Arizona, Arkansas, California, Colorado, Connecticut, Delaware, District of Columbia, Hawaii, Illinois, Iowa, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Jersey, New Mexico, New York, North Dakota, Ohio, Oregon, Rhode Island, Vermont, Washington, and West Virginia. Additional states could decide to expand their Medicaid programs at a later date subject to CMS approval. [39] 42 U.S.C. § 13031. [40] 42 U.S.C. § 18041. [41] 26 U.S.C. § 36B. [42] 42 U.S.C. § 18071. [43] For purposes of this study, we define the following occupations as blue-collar: mechanical/repair, construction trade/extractors, precision production, operators, and farming/forestry/fishing. For certain variables (e.g., employment and financial information), we captured responses from the survey wave in which the respondent was age 60 through 62 so we can compare early and delayed claimers at similar ages. We convert all dollar values to 2010 dollars. [44] GAO, Unemployed Older Workers: Many Experience Challenges Regaining Employment and Face Reduced Retirement Security, [hyperlink, http://www.gao.gov/products/GAO-12-445] (Washington, D.C.: Apr. 25, 2012). [45] Richard W. Johnson, Karen E. Smith, and Owen Haaga, "How Did the Great Recession Affect Social Security Claiming?" Urban Institute, Brief No. 37 (July 2013). Similarly, a separate study found that higher unemployment rates increase early claiming, mainly for workers with less education. See Courtney C. Coile & Phillip B. Levine, "Recessions, Retirement, and Social Security," American Economic Review: Papers & Proceedings, Vol. 101: 3, 2011. [46] As we previously reported, women who are divorced or separated experience higher rates of poverty. In particular, becoming divorced or separated after age 50 substantially reduces household assets and income, especially for women. See GAO, Retirement Security: Women Still Face Challenges, [hyperlink, http://www.gao.gov/products/GAO-12-699] (Washington, D.C.: July 19, 2012). [47] The total household wealth measure we use from the HRS is the net value of all wealth (including second home). [48] Experts have noted, however, that, if an individual is married, he or she should also base their decision on the life expectancy of their spouse to ensure that their survivor benefits are sufficient. [49] Hilary Waldron, "Do Early Retirees Die Early? Evidence from Three Independent Data Sets," Working Paper, S SA Office of Research, Evaluation, and Statistics, 2002. [50] About one-quarter of our sample reports they have a 90 percent or greater chance of living to age 75. [51] Hilary Waldron, "Trends in Mortality Differentials and Life Expectancy for Male Social Security-Covered Workers, by Socioeconomic Status," Social Security Bulletin, Vol. 67 No. 3, 2007. [52] S. Jay Olshansky, et. al., "Differences in Life Expectancy Due to Race and Educational Differences Are Widening, and Many May Not Catch Up," Health Affairs, 31, No. 8 (2012): 1803-1813. As noted earlier, our sample does not include those who receive Social Security Disability Insurance, which includes a higher share of blacks and those of another race compared to whites. [53] Olshansky, 2012. Ellen R. Meara, Seth Richards and David M. Cutler, "The Gap Gets Bigger: Changes in Mortality and Life Expectancy, by Education, 1981-2000", Health Affairs, 27, No. 2 (2008): 350-360. [54] Olshansky 2012. [55] "After claiming benefits" in this section refers to measurements of income and wealth taken in the first HRS wave in which the HRS respondent reports having income from Social Security benefits. The analysis does not adjust for cohort effects nor for the different ages for the different claiming group--those in the full retirement age or later claiming group will be older in the HRS wave after claiming than those in the before full retirement age claiming group. [56] The HRS includes not only income from "traditional" defined benefit plans in pension income, but also distributions from defined contribution plans. [57] In our analysis, "at age 72" refers to measures taken during the HRS wave in which the HRS respondent reports being 72 years old for the first time. [58] We do not, in this analysis, take into consideration any possible differences in mortality rates between early and late claimers from claiming until age 72 that might affect the composition of the two sub- samples. [59] For each income comparison group, we include households from each claiming group with total income within 20 percent of the benchmark income level (median, 10th-percentile, 90th-percentile) for the entire sample in the HRS wave after claiming benefits. For example, the median household income for the sample was $51,939; for our comparison near the median, we include households with total income in the wave following claiming benefits from $41,551 to $62,327. We compare these same households in the age-72 wave, regardless of their total income in this wave. [60] Note that for this analysis we take average incomes and wealth of households near the same total income level, as opposed to median income and wealth we previously reported for the entire sample of claimers. [61] For purposes of this objective, we examined health insurance coverage only for early claimers from age 62 through 64, too young to qualify for Medicare based on their age. Therefore, in this section of the report, the term "early claimers" is used differently than it is elsewhere in the report. [62] Some individuals may be eligible for coverage through programs, such as Medicaid, but not enroll. These individuals would be included in our estimates of early claimers not covered by government or employer-sponsored health insurance. [63] Federal law generally prohibits the imposition of any premiums and restricts cost-sharing for certain Medicaid beneficiaries, including for those below 100 percent FPL. For adults in this lowest income range, states cannot apply more than a nominal amount for cost- sharing, nor can they impose premiums or any cost-sharing for certain services, such as emergency services or hospice. States may choose, however, to impose somewhat higher cost-sharing charges for populations at higher income levels (such as between 100 and 150 percent FPL) for certain services, such as nonemergency use of the emergency room. [64] In addition to the plans available through the exchanges, individuals could choose to purchase individual coverage available outside of the exchanges. These plans are required to meet many of the same provisions as the plans available through the exchanges. However, premium tax credits and cost-sharing subsidies are only available for plans purchased through the exchanges. [65] These data include plan information from the 34 federally operated exchanges and the state-operated exchanges in Idaho and New Mexico. We did not examine the types of plans available in the other 15 state-operated exchanges, as comprehensive information was not readily available. [66] Within each state, we examined premiums for the rating area in which the state's most populous county was located. We focus our analysis on the second-lowest-cost silver plan because it is a mid- cost plan and is the type of plan that must be used to calculate individuals' premium tax credits. For comparison, on average, for a 62- year-old nonsmoker, the monthly premium for the lowest-cost bronze plan would be about $443, $140 less than for the second lowest cost silver plan. Premiums for the lowest-cost bronze plans ranged from $259 per month in Minnesota to $743 per month in Wyoming. [67] Plan premiums can vary by geographic area to account for geographic differences in cost of delivering care and the process for setting geographic rating areas and reviewing rates is left to individual states, subject to review by HHS. However, state capacity to conduct such reviews varies and HHS will conduct reviews in those states that lack the necessary resources or statutory authority. Starting in 2010, HHS began awarding grants to states to enhance their ability to review issuer rates. GAO, Patient Protection and Affordable Care Act: HHS's Process for Awarding and Overseeing Exchange and Rate Review Grants to States, [hyperlink, http://www.gao.gov/products/GAO-13-543] (Washington, D.C.: May 31, 2013). [68] In addition, some early claimers with the option for continuation coverage from their employer, as required under COBRA, could be eligible for tax credits. The Internal Revenue Service has proposed that COBRA coverage only would make an individual ineligible for the tax credits if that individual elects to enroll. Accordingly, assuming this rule is finalized, an individual could choose to discontinue their COBRA coverage and instead qualify for a premium tax credit for an exchange plan. [69] PPACA specifies the percentage of the reduction based on the individuals' income. For example, an individual with income between 100 and 133 percent FPL would be required to pay 2 percent of their income; individuals between 300 and 400 percent FPL would be required to pay 9.5 percent of their income. Advance payments of the premium tax credit are made by the federal government directly to the plan in which the individual enrolls. [70] Some of these totals do not add due to rounding. [71] If an individual chooses a more expensive plan than the second- lowest -cost -silver plan, the PPACA premium tax credit will not change and the individual will be responsible for the additional premium amount. If he or she chooses a plan with a less expensive premium than this plan, the individual will still be eligible to receive a PPACA premium tax credit up to the amount calculated based on the premium for the second-lowest-cost silver plan, which would reduce the amount of premium expenses the individual would have to pay. [72] In New York and Vermont early claimers will pay the same premiums as younger adults, as the states prohibit issuers from varying premiums by age. In these states, premiums for all consumers, including early claimers were $365 and $413, respectively. New York and Vermont are the only states that prohibit age-based rating. [73] According to the Henry J. Kaiser Family Foundation, the states that were not expanding Medicaid generally do not offer coverage to adults who are not disabled or without dependent children. These states do offer coverage to some adults with dependent children, but the median income eligibility level is 47 percent FPL. Henry J. Kaiser Family Foundation, The Coverage Gap: Uninsured Poor Adults in States that Do Not Expand Medicaid (Menlo Park, CA: October 2013). [74] According to one study, among employers offering retiree benefits, the median age of retirement in 2011 was 63; among employers without retiree coverage, the median age of retirement was 65. Mercer, LLC, National Survey of Employer-Sponsored Health Plans: 2012 Survey Report (New York, N.Y.: 2013). [75] The Congressional Budget Office (CBO) estimated that changes brought about by PPACA, including the implementation of provisions related to insurance coverage, will reduce the total number of hours Americans worked by 1.5 to 2.0 percent between 2017 and 2024, almost entirely due to individuals choosing to work less. CBO noted that one factor that could reduce the supply of labor is the decision of older workers to retire earlier than they otherwise would have due to the lower cost of health insurance available to them outside of the workplace and PPACA protections that prohibit denial of individual coverage based on health status. CBO, The Budget and Economic Outlook: 2014 to 2024 (Washington, D.C.: February 2014). [76] See GAO, Patient Protection and Affordable Care Act: Estimates of the Effect on the Prevalence of Employer-Sponsored Health Coverage, [hyperlink, http://www.gao.gov/products/GAO-12-768] (Washington, D.C.: July 13, 2012). [77] Mercer, LLC, National Survey of Employer-Sponsored Health Plans: 2011 Survey Report (New York, N.Y.: 2012). [78] Mercer, LLC, National Survey of Employer-Sponsored Health Plans: 2012 Survey Report (New York, N.Y.: 2013). [79] Public employees may or may not have Social Security coverage. State and local governments, in conjunction with the Social Security Administration (SSA), generally decide whether or not to provide Social Security coverage to their employees who are members of a public retirement system. [80] HRS Sample Evolution, University of Michigan, December 2008, [hyperlink, http://hrsonline.isr.umich.edu/sitedocs/surveydesign.pdf]. [81] HRS Sample Evolution, University of Michigan, December 2008, [hyperlink, http://hrsonline.isr.umich.edu/sitedocs/surveydesign.pdf]. [82] For many of the characteristics we analyze, we create dummy variables (e.g., a zero/one indicator of whether or not the respondent had less than a college degree). The reference category for any particular dummy variable may include respondents without the characteristics and missing values. [83] Odds (O) are mathematically related to, but not the same as, probabilities (P), that is O=P/[1-P]. [84] D. W. Hosmer, Jr. and S. Lemeshow., Applied Logistic Regression, 2nd ed. (New York: Wiley, 2000). [85] K. J. Archer, Goodness-of-fit tests for logisitic regression models developed using data collected from a complex sampling design, Ph.D. thesis (Ohio State University, 2000). [86] Kellie J. Archer and Stanley Lemeshow, "Goodness-of-fit test for a logistic regression model fitted using survey sample data " The Stata Journal, 6, Number 1 (2006): 97-105. [87] For this analysis we also eliminate all HRS respondents born prior to 1931 in order to limit cohort effects on benefits and on other sources of income that adjusting for inflation would not address. [88] Within the measure of pension income, RAND-HRS does not distinguish payments from a defined benefit plan, which tend to be paid as lifetime annuities, or a defined contribution plan, which would most likely represent withdrawals from an account. Similarly, it does not specify distributions from an IRA or other account-based retirement vehicle that could be exhausted. Capital income includes business or farm income, self-employment earnings, gross rent, dividend and interest income, trust funds or royalties, or other asset income. [89] To see a GAO report using a similar methodology for calculating household income, see GAO, Retirement Security: Women Still Face Challenges, [hyperlink, http://www.gao.gov/products/GAO-12-699] (Washington, D.C.: July 19, 2012). [90] That is, we take all households with total income within 20 percent of median income, not 20 percent of all households above and below the median (and similarly for households with income near the 10TH and 90TH income percentile level). [91] See Victoria Lynch et al., "Improving the Validity of the Medicaid/CHIP Estimates on the American Community Survey: The Role of Logical Coverage Edits" (Urban Institute Health Policy Center, Sept. 27, 2011). The researchers developed 16 additional edits applied to responses from children and 6 additional edits applied to responses from nonelderly adults, which when applied, resulting in a 3.5 and 0.8 percent increase in their Medicaid/CHIP estimates, respectively. [92] Steven Ruggles et al., Integrated Public Use Microdata Series: Version 5.0, machine-readable database (Minneapolis: University of Minnesota, 2010). [93] See State Health Access Data Assistance Center, "Defining 'Family' for Studies of Health Insurance Coverage," Issue Brief #27 (Minneapolis: University of Minnesota). [94] ACS's variable showing Social Security retirement income also includes individuals who receive Railroad Retirement income. In fiscal year 2013, the Railroad Retirement Board paid retirement-survivor benefits of about $11.6 billion to about 568,000 beneficiaries. In comparison, Social Security paid $816 billion to 58 million beneficiaries in 2013. [95] ACS asks respondents if they have coverage directly purchased from an insurance company, which would include coverage purchased through the individual market. However, it could also include certain types of group coverage. For example, it could include coverage purchased through an association. [96] The Indian Health Service's health coverage program, the contract health services program, is not considered health insurance and is not able to pay for all eligible health care services for eligible American Indians and Alaska Natives. In addition, individuals eligible to have their care paid through the contract health services program are expected to first try to obtain coverage from other sources, such as Medicaid or private health insurance. See GAO, Indian Health Service: Most American Indians and Alaska Natives Potentially Eligible for Expanded Health Coverage, but Action Needed to Increase Enrollment, [hyperlink, http://www.gao.gov/products/GAO-13-553] (Washington, D.C.: Sept. 5, 2013). [97] 26 U.S.C. § 36B, 42 U.S.C. § 1396a(e)(14). [98] For 2011, the federal poverty level (FPL) for a family of four was $22,350 in the lower 48 states and the District of Columbia, $27,940 in Alaska, and $25,710 in Hawaii. [99] PPACA also provides for a 5 percent income disregard when calculating income for determining Medicaid eligibility, which effectively increases this income level to 138 percent of the FPL. Two states established a higher income eligibility standard for Medicaid in 2014. The standards in the District of Columbia and Minnesota were 215 percent and 205 percent, respectively. [100] GAO, Indian Health Service: Most American Indians and Alaska Natives Potentially Eligible for Expanded Health Coverage, but Action Needed to Increase Enrollment, [hyperlink, http://www.gao.gov/products/GAO-13-553] (Washington, D.C.: Sept. 5, 2013). [101] HHS, Health Insurance Marketplace Premiums for 2014 Databook, downloaded October 30, 2013, [hyperlink, http://aspe.hhs.gov/health/reports/2013/marketplacepremiums/datasheet_ho me.cfm]. [102] Henry J. Kaiser Family Foundation, Subsidy Calculator: Premium Assistance for Coverage in Exchanges, accessed January 16, 2014, [hyperlink, http://kff.org/interactive/subsidy-calculator/]. [103] U.S. Census Bureau, Annual Estimates of the Resident Population: April 1, 2010 to July 1, 2012, accessed October 29, 2013, [hyperlink, http://www.census.gov/popest/data/counties/totals/2012/index.html]. [104] PPACA also provides for a 5 percent income disregard when calculating income for determining Medicaid eligibility, which effectively increases this income level to 138 percent of the FPL. Two states established a higher income eligibility standard for Medicaid in 2014. The standards in the District of Columbia and Minnesota were 215 percent and 205 percent, respectively. [End of section] GAO's Mission: The Government Accountability Office, the audit, evaluation, and investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through GAO's website [hyperlink, http://www.gao.gov]. Each weekday afternoon, GAO posts on its website newly released reports, testimony, and correspondence. To have GAO e-mail you a list of newly posted products, go to [hyperlink, http://www.gao.gov] and select "E-mail Updates." Order by Phone: The price of each GAO publication reflects GAO's actual cost of production and distribution and depends on the number of pages in the publication and whether the publication is printed in color or black and white. Pricing and ordering information is posted on GAO's website, [hyperlink, http://www.gao.gov/ordering.htm]. Place orders by calling (202) 512-6000, toll free (866) 801-7077, or TDD (202) 512-2537. Orders may be paid for using American Express, Discover Card, MasterCard, Visa, check, or money order. Call for additional information. Connect with GAO: Connect with GAO on facebook, flickr, twitter, and YouTube. Subscribe to our RSS Feeds or E mail Updates. Listen to our Podcasts. Visit GAO on the web at [hyperlink, http://www.gao.gov]. To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Website: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]; E-mail: fraudnet@gao.gov; Automated answering system: (800) 424-5454 or (202) 512-7470. Congressional Relations: Katherine Siggerud, Managing Director, siggerudk@gao.gov: (202) 512-4400: U.S. Government Accountability Office: 441 G Street NW, Room 7125: Washington, DC 20548. Public Affairs: Chuck Young, Managing Director, youngc1@gao.gov: (202) 512-4800: U.S. Government Accountability Office: 441 G Street NW, Room 7149: Washington, DC 20548. [End of document]