401(k) Savings. enefits. Benefits. Savings. Savin. Savings. Net Financial Assets. 401(k) Capitalized Value Sav. Contributions.

Transcription

1 AARP Public Policy Institute Employer-Provided Pensions: Less To Count On FREEDOM HEALTH FEAR RETIREMENT RELAXATION MENTORING HUNGER INSECURITY UNDERSAVING PRIVATION VOLUNTEERING COMMUNITY SECURITY INDEPENDENCE FAMILY FRUSTRATION WANT ANXIETY WORRY POVERTY enefits Contributions Tenure Worker Keog et financial assets ion Wealth 401(k) Benefits Social Security Education Non-discrimination tests Employer cial Security Savings Contributions Contributions Simple Contributions Final salary pension ker Saver s Keogh -discrimination tests Employer Credit Saver s Credit Tenure IRA Net Financial Assets Replacement Rate Net financial assets Contributions Saver s Credit Pension Wealth Hybrid pension plan ension WealthPension Wealth Tenure Contributions Pension Wealth Tenure Tenure nancial Assets curity Benefits Coverage Benefits Contributions TenureSaver s Worker Contributions Credit Contributions Tenur ion Tenure Wealth Social Security Saver s Tenure Final salary Worker Saver s pensioncredit Keogh Save Net financial assets s Contributions Benefits Non-discrimination tests nure Social Pension Wealth Saver s Credit Contributions Tenure Benefits Tenure Net financial assets Benefits Net Financial Net Assets Final salary pension Financial Social Assets Secur Saver s Credit Pension Wealth Benefits Net financial assets Net financial assets Net financial Social assets Security Tenure Saver Tenure Social Vested Replacement Employer Rate Security Net financial assets Pension Coverage Wealth Coverage Non-discrimination tests Non-discrimination tests Social efitssecurity Pension Net financial assets Coverage Benefits Worker Wealth Net Financial Worker Assets Keogh Replacement Hybrid Security Pension Wealth Benefit pension Rate plan Non-discrimination Non-discrimination tests Employer tests curity Contributions TenureSaver s Credit Non-discrimination Net Financial Keogh tests Assets Pension Non-discrimi We 401(k) Benefits Savings Monthly Income 401(k) Worker Net financial Net assets Replacement Rate Financ Final sets utions placement Rate Worker Hybrid Coverage Keogh pension plan Employer Worker Final alth Tenure Worker Keogh Benefits Savings Coverage Net financial Non-discrimination tests Contributions Employer Employer salary Saver s Coverage pension Cre Keogh assets Net financial assets Non-discrimination tests EmployerBenefits Contributions Coverage Net Financial 401(k) Net Financial Assets Saver s Net financial Contributions Credit assets Replacement nure ocial Saver s Credit Worker Security Contributions Savings 401(k) Keogh Benefits Saving Rate Pension 401(k) Replacem Hybrid pension planwealth Savings Final salary pension 401(k) 401(k) Pension Wealth Replacement Rate Keogh Savings Hybrid pen on-discrimination tests Final salary pension Hybrid pension plan n tests Employer Monthly Salary Social Income401(k) Worker ons Keogh Hybrid pension plan Security Coverage ancial Assets Benefits assets Net financial assets Wealth Capitalized Value Sav Replacem nure Hybrid Social pension Saver s Salary Security plan Replacement Rate Credit Coverage Final salary pension Final Hybrid salary Benefits ets Pension Wealth Vest Hybrid pens Net 401(k) pension Financial plan Final salary pension Employer enure Contributions Final salary pension Net financial assets Tenure AssetsHybrid pension Coverage plan Saver s Credit Coverage Coverage Net financial assets Non-discrimination tests Employer Coverage Benefits Capitalized 401(k) Keogh Value 401(k) Social Security Savin Keogh 401(k Ke

2 Employer-Provided Pensions: Less to Count On Sandy Mackenzie Ke Bin Wu AARP Public Policy Institute This Report was prepared under the direction of Janet Mc Cubbin. AARP s Public Policy Institute informs and stimulates public debate on the issues we face as we age. Through research, analysis and dialogue with the nation s leading experts, PPI promotes development of sound, creative policies to address our common need for economic security, health care, and quality of life. The views expressed herein are for information, debate, and discussion, and do not necessarily represent official policies of AARP. # October, , AARP. Reprinting with permission only. AARP, 601 E Street, NW, Washington, DC

3 Contents Employer-provided pensions: Less to Count On Pension Coverage of Workers over Time Coverage by Worker Characteristics Coverage by Type of Pension Coverage of Working Families Retirement Account Wealth Differences between 2004 and 2007 Implications of the Findings Conclusion Supporting Tables 1

4 Employer-Provided Pensions: Less to Count On This report assesses the breadth and quality of pension coverage in the United States. Pension coverage is low overall, and there are serious disparities of coverage by income, demographic, and firm characteristics. Certain segments of the population have virtually no coverage at all. Although the retirement pension that Social Security provides may suffice for many Americans, the lack of an adequately sized pension from former employers will create hardship for many others. Major findings and recommendations The coverage of the employer-provided pension system is limited, and there are marked disparities in coverage between rich and poor, Hispanic and white, well- and less welleducated and other social groups. The Automatic IRA would broaden coverage substantially, and improving the Saver s Credit would provide incentives for low- and moderate-income workers to save for retirement. Defined contribution (DC) plans are now much more prevalent than traditional defined benefit (DB) plans. Given the dominant role now played by 401(k)-type plans, it is important to broaden their coverage and to encourage workers to contribute as much as they can. Older Americans may be depriving themselves of valuable longevity insurance. Only a small fraction of 401(k) plan and IRA balances are being annuitized. Even distributions from DB plans are increasingly made in lump-sum form. Encouraging annuitization could be effective in increasing retirement income security. The report first examines the extent of pension coverage of workers and working families and the amount of retirement savings, with a focus on Next we highlight changes in coverage between 2004 and 2007, and explore the significance of the findings. An appendix includes detailed tables of pension coverage and retirement savings by demographic characteristics for workers, working families, and all families. 2 Employer-provided pensions: Less to Count On

5 Data Source and Definitions This report is an update of Coverage of Employer-Provided Pensions: Partial and Uncertain, published in August Its principal source is the Federal Reserve Board s triennial Survey of Consumer Finances (SCF) for 1989 through 2007 (the most recent year available). The SCF is a very rich source of information on household finances, and in particular retirement finances. Pension coverage can be measured in a number of ways. This report focuses on the coverage of individual workers, but also examines the coverage of working families (families with at least one working member) and the coverage of all families (which includes retirees and other nonworkers). It provides information on pension coverage for the head of a household and his or her spouse or unmarried partner, but not for other family members, whether they work or not. The report uses both a narrow and a broad definition of pension/retirement plan coverage. The narrow definition is based on coverage under an employer-provided pension plan. The broad definition is based on coverage by virtue either of ownership of an IRA/Keogh plan or membership in an employer-provided pension plan. The narrow and more traditional definition does not include workers who are not members of an employer-provided plan, whether or not they have an IRA/Keogh account. Under either definition, employees who are offered a plan but choose not to participate are not counted as covered. In this report we refer to all defined contribution plans and cash balance plans as 401(k)-type plans. Cash balance plans are legally defined benefit plans but they accrue an account for the employee much like a 401(k)-type plan. The report defines retirement account assets as the balances of individual retirement accounts and 401(k)-type plans. Our estimate of retirement saving does not include an estimate of the present value of the future stream of income that retirees with a DB plan can expect to receive. Ideally, the pension wealth associated with DB plans would be included; but, estimating this component of pension wealth is difficult. This said, the trend of declining membership in DB plans that the report has confirmed will almost certainly entail a decline in the value of DB plan wealth. 3

6 Pension Coverage of Workers over Time Overall coverage of the working population under either an employer-provided or individual retirement plan has remained flat for more than three decades. Coverage under this broad measure was 64 percent in 2007, which differs little from its 2004 level. Coverage under an employer-provided pension plan from the worker s current or any previous employment was 54 percent. Coverage under a pension from the current employer drops to only 50 percent, in line with the well-established tendency for the employer-provided pension system to cover about one in two workers (figure 1). The 401(k) plan (and other DC plans) continued to dominate the DB plan. Some 41 percent of the workforce is covered under a 401(k)-type plan, possibly in conjunction with a DB plan. Only 17 percent are covered by a DB plan, with or without a DC plan (figure 2), and only 9 percent are covered by a DB plan alone. The shift away from DB plans began in the early 1980s, and nearly all the decline between 1980 and the present day had taken place by A slight uptick in the coverage of workers by DB plans has occurred since 2004, but nothing dramatic. 1 Coverage by Worker Characteristics Coverage rates increase dramatically with a worker s age. The share of workers who have or have had coverage under some kind of employer-provided pension or retirement plan (IRA/Keogh) increases from 30 percent for workers aged to reach 77 percent for workers aged before beginning Figure 1 Flat as a Pancake: Percentage of Workers with Pension Coverage and Retirement Savings, 1989 to Percentage of Workers Year With IRA/Keogh With pension coverage from current job With pension coverage from any job With either pension or IRA/Keogh 1 Household surveys of pension plan participation like the SCF typically produce lower coverage estimates than employer surveys do. This difference does not, however, affect the trend. 4 Employer-provided pensions: Less to Count On

7 Figure 2 DC Dominates: Percentage of Workers with DB Coverage and with 401(k)-Type Coverage at Current Job, 1989 to Percentage of Workers Year With DB pension coverage from current job With 401(k)-type pension coverage from current job a decline (figure 3). Rates of coverage peak before retirement age is reached, probably reflecting the tendency of older workers with pensions or retirement plans to leave the labor force, which increases the share of workers without coverage. Coverage rates are also strongly related to salary. Only 34 percent of workers paid less than 20,000 have pension or retirement plan coverage, compared with 90 percent of those earning 100,000 or more (figure 4). The relationship between the rate of coverage and salary may reflect the fact that low salary earners tend to be relatively young, and they may be less interested in a pension plan and a tendency for employers to devote pension benefits to their betterpaid staff. 2 Given the strong association between wage levels and the rate of pension coverage, it is not surprising to find a strong relationship between educational level and the rate of coverage. Coverage is very low for workers lacking a high school diploma, but increases substantially for those with a diploma (figure 5). Attending college without earning a degree does not increase coverage much, but earning a college degree does. Pension coverage also varies markedly from one ethnic group to another. Hispanic workers have the lowest rates and African Americans are next. The rates of whites and Asians and other ethnic groups are almost equal (see figure 6). Part of the observable differences in the rates of coverage of ethnic 2 Membership in 401(k) plans is optional, and younger workers are more likely to decline to participate. See figure 21 in Employment-based retirement plan participation: Geographic differences and trends, 2007, EBRI Issue Brief No. 322 (October 2008). 5

8 Figure 3 Percentage of Workers with Pension Coverage and Retirement Savings by Age, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Age 15 to 24 Age 25 to 34 Age 35 to 49 Age 50 to 59 Age 60 to 64 Age 65+ All workers Age Figure 4 Percentage of Workers with Pension Coverage and Retirement Savings by Annual Wage or Salary, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Under 20,000 20,000 to 29,999 30,000 to 44,999 45,000 to 59,999 60,000 to 74,999 75,000 to 99, ,000+ Annual Wage/Salary 6 Employer-provided pensions: Less to Count On

9 Figure 5 Percentage of Workers with Pension Coverage and Retirement Savings by Education, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh No high school diploma High school diploma Education Some college College or advanced degree Figure 6 Percentage of Workers with Pension Coverage and Retirement Savings by Race/Ethnicity, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh White African American Hispanic Asian and Others Race/Ethnicity 7

10 groups may be explained by differences in educational attainment and average age. There is little difference in coverage rates by sex (see figure 7). The coverage of widowed workers and divorced or separated workers is higher than the coverage of workers who are married or who have never been married (see figure 8). Pension coverage is substantially affected by whether a worker is part-time or fulltime. The gap in coverage under the broad definition is less pronounced than it is under the narrow definition, perhaps because some part-timers are using the IRA/Keogh to compensate for their non-participation in employer-provided plans (see figure 9). Coverage by Type of Pension In general, the patterns observed with the broad measure of coverage are replicated for the narrower measure (employer plans) and for each of the two major employer plan types. Coverage under both defined benefit plans and 401(k)-type plans rises substantially with age and salary (see figures 10 and 11). In light of the low overall coverage of defined benefit plans, it is not surprising that they cover no more than one in twenty young workers or workers with low salaries. Although coverage for each type of pension by ethnic group has a pattern similar to overall coverage, the defined benefit coverage of whites is slightly higher than that of Asians and others (see figure 12). Coverage by sex is similar under either plan type (see figure 13). As for marital status, widows or widowers are more likely to be covered by a defined benefit plan than Figure 7 Percentage of Workers with Pension Coverage and Retirement Savings by Sex, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Men Sex Women 8 Employer-provided pensions: Less to Count On

11 Figure 8 Percentage of Workers with Pension Coverage and Retirement Savings by Marital Status, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Married Never married Widowed Divorced or separated Marital Status Figure 9 Percentage of Workers with Pension Coverage and Retirement Savings by Full-Time and Part-Time Status, Percentage of Workers With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Full-Time Part-Time Full-Time/Part-Time Status 9

12 Figure10 Percentage of Workers with Pension Coverage at Current Job by Age, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Age 15 to 24 Age 25 to 34 Age 35 to 49 Age 50 to 59 Age 60 to 64 Age 65+ Age Figure 11 Percentage of Workers with Pension Coverage at Current Job by Annual Wage/Salary, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Under 20,000 20,000 to 29,999 30,000 to 44,999 45,000 to 59,999 Annual Wage/Salary 10 Employer-provided pensions: Less to Count On 60,000 to 74,999 75,000 to 99, ,000+

13 Figure 12 Percentage of Workers with Pension Coverage at Current Job by Race/Ethnicity, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type White African American Hispanic Asian and Others Race/Ethnicity Figure 13 Percentage of Workers with Pension Coverage from Current Job by Sex, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Men Sex Women 11

14 are married workers, perhaps because the former group is on average older and more likely to have worked in an industry where defined benefit plans were commonplace (see figure 14). Firm size is strongly related to the rate of coverage; the rate of coverage for workers at establishments employing fewer than 20 people is only 12 percent. For workers at establishments employing more than 500 people, coverage rises to 72 percent (see figure 15). Economies of scale in plan administration would be one influence on the relationship between firm size and coverage. In addition, large firms pay their employees better. 3 Coverage also varies strongly with full-time/part-time status; hardly any part-time workers are covered by a DB plan (see figure 16). Tenure (length of time on the job) and pension coverage are also strongly related. DB coverage approaches 30 percent with 15 or more years of tenure (see figure 17). Tenure is associated with age. The longer their average tenure, the older are workers likely to be, and the more likely to be covered by a DB plan. In addition, continued participation in a DB plan is more attractive the longer a plan member s period of service. Coverage of Working Families Measuring coverage on a per-family basis captures the role of the family as a safety net. A worker without coverage through Figure 14 Percentage of Workers with Pension Coverage at Current Job by Marital Status, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Married Never married Marital Status Widows Divorced or separated 3 The Bureau of Labor Statistics release Employer costs for employee compensation June 2009 finds that total compensation per hour by establishment size is as follows: 1 49 workers, 26.02; workers, 29.99; workers, 31.63; 500 or more workers, Employer-provided pensions: Less to Count On

15 Figure 15 Percentage of Workers with Pension Coverage at Current Job by Firm Size, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Less than 10 employees 10 to 19 employees 20 to 99 employees Firm Size 100 to 499 employees 500+ employees Figure 16 Percentage of Workers with Pension Coverage at Current Job by Full-Time and Part-Time Status, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Full-Time Full-Time/Part-Time Status Part-Time 13

16 Figure17 Percentage of Workers with Pension Coverage at Current Job by Tenure, Percentage of Workers DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Less than 2 years 2 to 4 years 5 to 9 years 10 to 14 years Years Working for Current Employer 15 to 19 years 20 years or more his or her own employer may receive pension benefits through his or her spouse s participation in a pension or retirement plan. However, coverage by working family is not markedly higher than coverage per worker. In particular, 71 percent of working families have either a pension or a retirement saving plan, which is 7 percentage points above the estimate for coverage by worker (see figure 18). As expected, coverage of working families increases by age and family income (see figures 19 and 20.) Working families are much more likely than individual workers to be covered by a defined benefit plan (24.3 percent versus 16.6 percent for coverage on the current job), particularly at older ages and higher incomes. This may reflect a tendency for workers with DB pensions to be paired with workers without coverage (see figures 21 and 22). Retirement Account Wealth Retirement plan balances are typically low the median value of the sum of IRA and 401(k)-type plan holdings in 2007 among working families with retirement savings was estimated to amount to only 46,000 and, as the substantial difference between median and mean values attest, retirement wealth is highly concentrated (figure 23). As might be expected, both the distribution by age and by income quintile are also highly concentrated, although holdings of working families whose head is 65 or older are lower than those of the group aged (figures 24 and 25). This difference might reflect a tendency for older cohorts to rely more 14 Employer-provided pensions: Less to Count On

17 Figure 18 Percentage of Working Families with Pension Coverage and Retirement Savings, 2007 With neither pension nor IRA/Keogh With pension only 29.3% 37.4% With both pension and IRA/Keogh % With IRA/Keogh only Figure19 Percentage of Working Families with Pension Coverage and Retirement Savings by Age of Family Head, Percentage of Working Families With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Age 15 to 24 Age 25 to 34 Age 35 to 49 Age 50 to 59 Age 60 to 64 Age 65+ Age of Family Head 15

18 Figure 20 Percentage of Working Families with Pension Coverage and Retirement Savings by Family Income, Percentage of Working Families With pension coverage from current job With pension coverage from any job With IRA/Keogh With either pension or IRA/Keogh Under 20,000 20,000 to 29,999 30,000 to 44,999 45,000 to 59,999 Family Income 60,000 to 74,999 75,000 to 99, ,000+ Figure 21 Percentage of Working Families with Pension Coverage at Current Job by Age of Family Head, Percentage of Working Families DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Age 15 to 24 Age 25 to 34 Age 35 to 49 Age 50 to 59 Age 60 to 64 Age 65+ Age of Family Head 16 Employer-provided pensions: Less to Count On

19 Figure 22 Percentage of Working Families with Pension Coverage at Current Job by Family Income, Percentage of Working Families DB 401(k)-type Both DB and 401(k)-type Either DB or 401(k)-type Under 20,000 20,000 to 29,999 30,000 to 44,999 45,000 to 59,999 Family Income 60,000 to 74,999 75,000 to 99, ,000+ Figure 23 Mean and Median Retirement Saving Assets, 2007 (Working Families with Retirement Savings Only) 160, ,000 Retirement Saving Assets Mean Median 141, , , , ,951 80,000 60,000 40,000 30,000 32,500 46,000 20,000 0 IRA/Keogh balance 401(k)-type balance Total retirement saving assets 17

20 Figure 24 Median Retirement Saving Assets by Age of Family Head, 2007 (Working Families with Retirement Savings Only) Median Retirement Saving Assets (thousands) IRA/Keogh assets 401(k)-type pension assets Total retirement saving assets Age 15 to 24 Age 25 to 34 Age 35 to 49 Age 50 to 59 Age 60 to 64 Age 65 and older Age of Family Head Figure 25 Median Retirement Saving Assets by Family Income Quintile, 2007 (Working Families with Retirement Savings Only) Median Retirement Saving Assets (thousands) IRA/Keogh assets 401(k)-type pension assets Total retirement saving assets Lowest Quintile Second Quintile Third Quintile Family Income Quintile 18 Employer-provided pensions: Less to Count On Fourth Quintile Highest Quintile

21 on pensions, especially DB pensions, than on retirement saving plans. It could also reflect some degree of spending down by the older cohort. Savings in retirement plans and even net financial assets by families with no pension are very low, and cannot compensate them for the lack of pension coverage (figure 26). The holdings of families with pension coverage are much higher than those of families without. Differences between 2004 and 2007 The main indicators of pension and retirement saving plan coverage have changed little between 2004 and 2007, as is evident from figure 1. Coverage of workers measured broadly (i.e., including both employer-provided pensions and retirement saving plans) rose only slightly, from 63 percent in 2004 to 64 percent in Coverage from an employer-provided pension alone remains just over 50 percent (53.5 percent in 2007 and 52.3 percent in 2004). The share of defined benefit pensions is not very different from what it was in A similar constancy can be observed in the coverage rates of working families, which continue to exceed the coverage rates of workers in the same age bracket. The 2007 survey does not suggest that fundamental changes in the trend of pension coverage are taking place, but there are some interesting developments to report. The latest survey confirms the convergence of the rate of coverage of women in the workforce by an employer-provided with that of men, which appears to have ended a long period when the rate of coverage of men was higher than that of women (see figure 27). Figure 26 Median Retirement Saving Assets and Net Financial Assets for Working Families with and without Pension Coverage, 2007 (Working Families with Retirement Savings and Net Financial Assets Only) 120, , ,000 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 Median Assets 104,600 50,000 With Pension Retirement savings Net financial assets 27,000 21,400 Without Pension 19

22 Figure 27 Percentage of Workers with either DB or 401(k)-Type Pension Coverage in Current Job by Sex, 1989 to Percentage of Workers Year Men Women The rate of coverage of Asian and other non-white, non-african-american, and non-hispanic workers, whether broadly or narrowly defined, which has been on an increasing trend for some time, exceeded that of white workers for the first time in 2007 and accounts for much of the overall increase already noted (figure 28). In contrast to these marginal changes, there has been little change in the pattern of coverage by full-time/part-time status, firm size, or tenure. There has been a slight decline in the coverage rates for workers at firms with fewer than 20 employees, and the rate of coverage of employer-provided pensions declined for workers in the age group and increased in the 65 and older group. As for the amount of retirement savings, the distribution of assets in retirement plans across age and income groups has changed little (table 1). The distribution of retirement savings remains highly skewed, with the top 20 percent of working families holding twothirds of retirement assets. Implications of the Findings Role of the DB pension. The traditional DB pension has been in decline for some time. Most of that decline took place before 1995, and the additional decline in coverage after that date has been limited. Although the slight uptick in DB coverage for workers in 2007 could be a sign that the decline has leveled off, the size of the increase is too modest to warrant a definite conclusion. 20 Employer-provided pensions: Less to Count On

23 Figure 28 Percentage of Workers with either DB or 401(k)-Type Pension Coverage in Current Job by Race/Ethnicity, 1989 to Percentage of Workers Year White African American Hispanic Asian and Other Table 1: Distribution of Retirement Assets of Working Families by Income Quintile and Age, 2004 and Family Income Quintile Mean holdings (dollars) with coverage (thousands) Share of total assets (percent) First quintile 17,268 20,736 3,076 4, Second quintile 29,384 33,107 7,049 7, Third quintile 46,050 64,822 10,063 10, Fourth quintile 83, ,406 12,322 12, Fifth quintile 268, ,379 13,795 14, Age of Family Head 15 to 24 9,707 9,529 1,174 1, to 34 28,755 28,563 8,226 8, to 49 85, ,841 19,508 20, to , ,605 11,435 13, to , ,836 3,518 3, and older 250, ,211 2,444 2, Working families with retirement savings only. Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, 2004 and

24 The relationships between age, salary, education, and other variables and pension coverage. However it is measured, pension coverage is associated with age, salary, education, ethnicity, and other characteristics. However, the associations between one variable or trait and the rate of coverage do not allow us to say much about the actual separate contribution of that trait. For example, does an increase in a worker s age, other things equal, increase the likelihood he or she will be covered, or is the apparent relationship between age and coverage merely capturing the well-known relationship between age and average salary? Similarly, does education exercise an independent influence on the rate of coverage, or does it merely reflect the fact that income tends to rise with educational attainment? Similar observations can be made about the influence of other variables. Part-time work tends to be less well paid than full-time work; at the same time, the cost of providing a pension to a part-time worker is greater than the cost of providing a pension to a full-time worker. Tenure could be related to the rate of coverage because long-tenured workers are older and better paid, because older workers are more likely to have been members of defined benefit plans, or because the longer a worker s tenure, the less the incentive to shift to a new plan. A forthcoming study by AARP will try to discern the separate influences of these variables, and thereby provide a more precise answer to the question of the determinants of pension plan coverage. Parity of rates of coverage of men and women. This is a heartening development. Nonetheless, the rate of coverage of neither sex is high. In addition, because women are much more likely to be covered by a 401(k)-type plan than a DB plan, their greater longevity means that they must save more than their male counterparts to receive the same steady lifelong income in retirement. How high should the rate of coverage be? The typical American worker will not be able to rely on the Social Security retirement benefit alone for income in retirement. The Social Security benefit can provide a fairly high replacement rate for the low and moderately paid. However, the replacement rate drops off quite sharply as income rises. 4 In principle, a worker could adequately supplement his or her Social Security benefit by taking advantage of the IRA/Keogh, which is tax favored, or even by saving that is not tax favored. The evidence on actual holdings suggests that this is not happening, and that fewer than one in ten eligible taxpayers contribute to an IRA each year. 5 The evidence suggests that retirement security for many will depend on the broad coverage of employer-provided pensions with adequate benefits. But how broad and how adequate? Rates of coverage for young and modestly paid workers are clearly low. Some 4 In 2009, the marginal replacement rate is 90 percent for the first 744 of average indexed monthly earnings, 32 percent for the next 3,739, and 15 percent on any monthly earnings higher than 4, See Bryant, Victoria L., Accumulation and Distribution of Individual Retirement Arrangements, 2004, Statistics of Income Bulletin, Spring 2008, Internal Revenue Service. 22 Employer-provided pensions: Less to Count On

25 part of this could be explained by the need of younger workers to accumulate enough equity to buy a house. It might also reflect a decision to put off saving until increases in income make it less of a sacrifice. This said, the rate of coverage of employerprovided plans is never very high, except for the best-paid workers. The rate of coverage from a pension provided by the current or any past employer increases with age, but does not rise above 60 percent until the age group. Rates of coverage in and of themselves do not allow us to determine whether an employer-provided pension system will generate an adequate contribution to retirement security. The increased importance of the 401(k) plan may not be contributing significantly to the adequacy of pensions even among those workers who remain covered. First, the actual rate of saving may be too low to generate adequate growth in account balances. The median retirement saving balance for working families in the age group is 106,000. This sum would buy an inflation-indexed annuity of no more than 4,000 per year, and half of the age group has asset holdings that do not match this figure. Second, the share of the balance of the plans of many older Americans invested in stocks remains high. 6 Finally, the balances of 401(k) plans and IRAs are seldom used to purchase an annuity. 7 Remedies. Several initiatives, either in place or contemplated, could increase coverage and enhance retirement security for the more vulnerable of America s workers. These include auto-enrollment in 401(k) plans, which many large companies have already put in place. There is considerable scope for boosting 401(k) coverage rates, since about 20 percent of workers whose employers provide these plans do not join them. 8 Changes to pension law introduced in the Pension Protection Act of 2006 have clarified and modified the law relating to automatic enrollment, and final regulations issued by the Department of Labor on Qualified Default Investment Alternatives have encouraged employers to adopt an auto-enrollment feature. These measures can be expected to reduce the rate of non-participation. In addition, the use of default contribution settings can be used to increase the number of workers who take full advantage of their employer s match. Auto-escalation an increase in the rate of contribution to 401(k) plans that takes place automatically unless the employee specifically declines it is already a feature of many DC plans, and its further spread would boost retirement saving. Measures to promote sound asset management and annuitization (such as financial education or allowing for a trial annuitization period in which a retiree can 6 The Employee Benefit Research Institute (EBRI) estimates that 401(k) members in their 60s who did not hold company stock held 47 percent of their balance in equity funds at end-year 2007 (EBRI, 401(k) Plan Asset Allocation, Account Balances, and Loan Activity, in 2007 Issue Brief No. 324 (December 2008)). 7 Of employers surveyed by Hewitt Associates who offered a 401(k) plan, only 15 percent offered an annuity option, and only 6 percent of retirees offered that option chose to take it. (EBRI Notes, Vol. 29, No. 11 (November 2008)). 8 This is the estimate made by the Profit Sharing Council of America (PCSA) for plans in See PSCA, 51st Annual Survey of Profit Sharing and 401(k) Plans. Other authorities have made similar estimates. 23

26 learn about the annuity) would increase retirement income security for workers who do not receive a traditional, annuitized, DB benefit. An important proposed initiative is the automatic IRA. The automatic IRA would require all employers with 10 or more employees who do not currently offer a retirement plan to offer a payroll deduction Individual Retirement Account. In addition to expanding access to a workplace retirement plan, the automatic IRA proposal would significantly increase participation by automatically enrolling workers who do not elect to opt out so that no workers are left out because of inertia. The adoption of an auto IRA would make substantial inroads among small employers, for whom more conventional forms of pension are too costly. The Saver s Credit should be modified to encourage more low- and moderateincome taxpayers more effectively to save for retirement. The Saver s Credit is a nonrefundable tax credit available to lowand moderate-income taxpayers contributing to a 401(k)-type plan or IRA. The credit is structured to offer lower-income workers a greater incentive to participate in a pension plan than higher-income workers; this structure is the opposite of the existing incentive structure of taxable income exclusions and deductions in which those with the highest income receive the greatest benefit. Potential improvements include making the tax credit refundable, making the current three-tier credit rate structure simpler, and making the credit more like an employer match by depositing the tax credit in the retirement account. Conclusion This report confirms for 2007 our findings for 2004: however pension coverage is measured, it remains low, just over 50 percent. Coverage rates vary markedly by age group and income level and with ethnicity, part-time versus full-time status, size of plan sponsor, and tenure. The low rates of coverage of the employer-provided pension system remain a matter of real concern. Workers without coverage through an employer are not making up the difference through contributions to IRAs or other personal saving. Although the basic picture for 2007 does not differ much from earlier assessments, there are some interesting differences. The gap in rates of coverage between women and men is now very small, and coverage of Asians has exceeded that of whites for the first time. Even if there are good reasons for rates of coverage of employer-provided plans to be lower among the young and the low-paid than they are among the older and better paid, rates of coverage among workers in the prime of life and earning a decent salary are sufficiently low to raise serious doubts about retirement security. Immediate action to increase low coverage rates among younger and less well paid workers and perhaps others should be a policy priority. 24 Employer-provided pensions: Less to Count On

27 Supporting Tables Table A.1: Workers with Pension Coverage and Retirement Savings by Sex, Age, and Race/Ethnicity, 2007 With Pension Coverage from Current Job With Pension Coverage from any Job With IRA/Keogh With either Pension or IRA/Keogh Without any Pension or IRA/Keogh Total of Workers Total Workers 53, , , , , ,773 Sex Men 28, , , , , ,059 Women 24, , , , , ,714 Age 15 to 24 1, , , , , to 34 9, , , , , , to 49 22, , , , , , to 59 15, , , , , , to 64 2, , , , , , and older 1, , , , , ,668 Race/Ethnicity White 40, , , , , ,529 African American 5, , , , , ,633 Hispanic 3, , , , ,407 Asian and Others 3, , , , , ,205 Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances,

28 Table A.2: Workers with Pension Coverage and Retirement Savings by Education, Annual Wage/Salary, Full-time/Part-time Status and Marital Status, 2007 With Pension Coverage from Current Job With Pension Coverage from any Job With IRA/Keogh With either Pension or IRA/Keogh Without any Pension or IRA/Keogh Total of Workers Total Workers 53, , , , , ,773 Education No High School Diploma High School Diploma Only 1, , , , ,477 13, , , , , ,339 Some College 9, , , , , ,320 College Degree 28, , , , , ,638 Annual Wage/Salary Less than 20,000 3, , , , , ,537 20,000 to 6, , , , , ,422 29,999 30,000 to 13, , , , , ,664 44,999 45,000 to 10, , , , , ,572 59,999 60,000 to 6, , , , , ,144 74,999 75,000 to 6, , , , , ,460 99, ,000 or more 7, , , , , ,973 Full- or Part-Time Status Full-time 51, , , , , ,413 Part-time 1, , , , , ,360 Marital Status Married 41, , , , , ,145 Never Married 4, , , , , ,798 Widowed , ,536 Divorced or Separated 6, , , , , ,295 Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, Employer-provided pensions: Less to Count On

29 Table A.3: Workers with Pension Coverage at Current Job by Sex, Age, Race/Ethnicity, and Annual Wage/Salary, 2007 DB 401(k)-Type Both DB and 401(k)-Type Either DB or 401(k)-Type Total Workers 17, , , , Sex Men 9, , , , Women 8, , , , Age 15 to , , to 34 2, , , to 49 7, , , , to 59 5, , , , to 64 1, , , and older , , Race/Ethnicity White 13, , , , African American 1, , , Hispanic , , Asian and Others , , Annual Wage/Salary Less than 20, , , ,000 to 29,999 1, , , ,000 to 44,999 4, , , , ,000 to 59,999 3, , , , ,000 to 74,999 2, , , , ,000 to 99,999 2, , , , ,000 or more 2, , , , Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, % 27

30 Table A.4: Workers with Pension Coverage at Current Job by Worker s Tenure, Firm Size, Full-time/Part-time Status and Marital Status, 2007 DB 401(k)-Type Both DB and 401(k)-Type Either DB or 401(k)-Type Total Workers 17, , , , Tenure Less than 2 years 1, , , to 4 years 2, , , , to 9 years 3, , , , to 14 years 2, , , , to 19 years 2, , , , years or more 4, , , , Firm Size Less then 10 employees , , to 19 employees , , to 99 employees 1, , , to 499 employees 3, , , , or more employees 12, , , , Full- or Part-Time Status Full Time 17, , , , Part Time , , Marital Status Married 13, , , , Never Married 1, , , Widowed Divorced or Separated 2, , , , Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, % 28 Employer-provided pensions: Less to Count On

31 Table A.5: Workers with Pension Coverage at Current or Previous Job by Sex, Age, Race/Ethnicity, and Annual Wage/Salary, 2007 DB 401(k)-Type Both DB and 401(k)-Type Either DB or 401(k)-Type Total Workers 22, , , , Sex Men 12, , , , Women 9, , , , Age 15 to , , to 34 2, , , to 49 9, , , , to 59 7, , , , to 64 2, , , and older , , Race/Ethnicity White 18, , , , African American 2, , , , Hispanic 1, , , Asian and Others , , Annual Wage/Salary Less than 20,000 2, , , ,000 to 29,999 2, , , ,000 to 44,999 4, , , , ,000 to 59,999 4, , , , ,000 to 74,999 2, , , , ,000 to 99,999 3, , , , ,000 or more 3, , , , Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, % 29

32 Table A.6: Working Families with Pension Coverage and Retirement Savings by Age of Family Head, Marital Status, and Family Income, 2007 With Pension Coverage from Current Job With Pension Coverage from any Job With IRA/Keogh With either Pension or IRA/Keogh Without any Pension or IRA/Keogh Total Working Families Total Workers 46, , , , , ,735 Age 15 to 24 1, , , , , to 34 8, , , , , , to 49 18, , , , , , to 59 12, , , , , , to 64 3, , , , , and older 2, , , , ,254 Marital Status Married Couple 34, , , , , ,106 Never Married 5, , , , , ,798 Widowed , ,536 Divorced or Separated Family Income 1 Less than 20,000 6, , , , , , , , , ,464 20,000 to 2, , , , , ,259 29,999 30,000 to 6, , , , , ,363 44,999 45,000 to 6, , , , , ,947 59,999 60,000 to 6, , , , , ,038 74,999 75,000 to 99,999 7, , , , , , ,000 or more 16, , , , , ,387 1 Income in the survey is reported for 2006 and is expressed here in 2007 dollars. Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, Employer-provided pensions: Less to Count On

33 Table A.7: Working Families with Pension Coverage at Current Job by Age of Family Head, Marital Status, and Family Income, 2007 DB 401 (k)-type Both DB and 401(k)-Type Either DB or 401(k)-Type Total Working Families 19, , , , Age 15 to , , to 34 2, , , to 49 7, , , , to 59 5, , , , to 64 2, , , and older 1, , , Marital Status Married Couple 14, , , , Never Married 1, , , Widowed Divorced or Separated 2, , , , Family Income 1 Less than 20, ,000 to 29, , , ,000 to 44,999 2, , , ,000 to 59,999 2, , , , ,000 to 74,999 2, , , , ,000 to 99,999 3, , , , ,000 or more 7, , , , Income in the survey is reported for 2006 and is expressed here in 2007 dollars. Data Source: AARP Public Policy Institute tabulations of the Federal Reserve Board s Survey of Consumer Finances, % 31