EBRI Retirement Income Research: 2004 Findings

Retirement Confidence Survey

From the April 2004 Issue Brief, "Will Americans Ever Become Savers? The 14th Retirement Confidence Survey, 2004."

  • This Issue Brief reports on the findings from the 14th annual Retirement Confidence Survey (RCS), a comprehensive study of the attitudes and behaviors of American workers and retirees toward saving, retirement planning, and long-term financial security. The 2004 RCS interviewed 1,002 individuals (783 workers and 219 retirees) age 25 and older in the U.S. and found that key attitudes and behavior patterns about retirement have barely changed over the past few years. The RCS was first released in 1991 and is unique in its long-term ability to track public attitudes about saving and retirement.
  • Many American workers have saved some money for retirement, but the proportion of Americans saying they have saved money for retirement has not increased since 2001. Moreover, 4 in 10 workers say they are not currently saving for retirement. Many of those with savings cite low levels of savings and investments.
  • Some workers may have expectations about their retirement that cannot be achieved. They expect to be able to work until they reach the age at which they want to retire, and a majority expect to be able to work for pay in retirement. Some will find that ill health or job loss ruins these plans. Workers also tend to expect their financial standard of living in retirement to be at least as good as before retirement and to remain at least at that level throughout the length of their retirement. For some this is unlikely, due to increasing medical costs, declining savings, and inflation.
  • Some workers may be counting on employer-provided retirement benefits that they may never receive. Workers are as likely to expect to receive benefits from a defined benefit plan pension and retiree medical insurance as current retirees are to have actually received them. At the same time, employers are cutting back on the provision of these benefits, making it less likely that future retirees will receive them. In some cases, workers who do receive these types of benefits will find that they are not as rich as those received by retirees in the past.
  • Few workers appear to have an idea of how much it takes to live comfortably in retirement. Only about 4 in 10 have taken steps to calculate how much they need to save by the time they retire in order to live comfortably in retirement, and one-third of those say they don't know or can't remember the result of the calculation.
  • Almost half of workers who have not saved for retirement feel at least some confidence about their ability to have a comfortable retirement. Some of these workers expect an employer to fund their retirement. Others are planning to save later, to rely on Social Security, to obtain support from family and friends, to work in retirement, or to manage through some other arrangement.
  • Retirement education can lead to changes in savings behavior of a significant proportion of workers. More than 4 in 10 workers who tried to do a savings needs calculation report changing their retirement planning as a result. Similarly, almost 3 in 10 of those who received retirement education through the work place changed their retirement planning.

Survey of Consumer Finance

From the January 2004 Notes article, "Retirement Plan Participation and Asset Allocation"

  • EBRI analysis of the Survey of Consumer Finance (SCF) showed there was a significant increase in the percentage of family heads with a defined contribution plan. In 1992, 57.8 percent of family heads participated in a defined contribution plans. That percentage increased to 78.7 percent in 2001.
  • The percentage of family heads that participated in a defined benefit plan declined over the same time period. In 1992, 59.3 percent of family heads participated in a defined benefit plans that percentage declined to 38.4 percent in 2001.
  • Overall, the percentage of family heads that participated in any type of employment-based retirement plan increased slightly from 48.3 percent in 1992 to 48.7 percent in 2001.
  • SCF classifies where IRA and 401(k)-type plan owners hold their assets in broad general asset categories: 1) mostly or all in stocks, 2) mostly or all in interest-earning assets, 3) split between stock and interest-earning assets, 4) other, although there are small difference in the classifications between the two types of accounts. Thus, the results do not give exact percentages, but they provide general insight into how individuals allocate their dollars within these plan types.
  • Among IRA participants, 53.5 percent had their assets invested in mostly stocks in 2001, while 401(k)-type plan participants had a similar percentage of 52.7 percent mostly invested in stocks. This is a substantial increase for both participant types: from 32.0 percent and 21.1 percent respectively, in 1992.

EBRI-ERF Retirement Security Projection Model (RSPM) and Retirement Income Projection Model (RIPM)

Future Retirement Income

From the February Issue Brief, "Americans' Future Retirement Security: Implications of the EBRI-ERF Retirement Security Projection Model."

  • This Issue Brief briefly summarizes the initial results from the national version of the Retirement Security Projection Model (reported in greater detail in the November 2003 EBRI Issue Brief), and provides analysis and comments on the implications of these findings from numerous professionals with policy expertise on the issues involved.
  • The United States faces the growing societal problem of a dramatic increase in the proportion of retirees relative to workers, and how to finance their care, due to such well-documented demographic trends as the largest generation (the baby boomers) nearing retirement, longer average life-spans, and far greater time spent in retirement.
  • According to a recent EBRI analysis done in collaboration with the Milbank Memorial Fund, the total annual national retirement income shortfall will grow to as much as $57 billion by 2030, an increase of approximately $25 billion from its projected level in 2005. During the decade ending in 2030, aggregate retiree income shortfall could exceed $500 billion.
  • The results show that for the median individual in birth cohorts on the verge of retirement there is little possibility of saving enough to supplement the simulated retirement wealth to provide adequate retirement income to meet basic needs. Younger birth cohorts would benefit from the increased years of contributions and would have savings targets that are feasible for most groups. However, there are some notable exceptions: Single females in the lowest income quartile are predicted to need in excess of 25 percent of compensation per year to have sufficient retirement wealth regardless of birth cohort-an improbably high goal.
  • An important issue in retirement security is large unanticipated health care costs-particularly if retirees need extended nursing home care or extended home health care. Furthermore, fewer retirees are going to have health insurance from a former employer and Medicare is projected to be severely underfunded once the baby boom generation stars retiring. All of this indicates more insecurity for future retirees about their ability to cover health expenditures in retirement.
  • Besides the personal implications for individual Americans who are likely to outlive their assets, the increasing national retirement income gap is expected to put heavy stress on state programs, particularly Medicaid, that finance long-term care for indigent retirees. However, because of budget difficulties, states are cutting spending on such programs rather than expanding them.

From the May 2004 Issue Brief, "ERISA at 30: The Decline of Private-Sector Defined Benefit Promises and Annuity Payments. What Will it Mean?"

  • This analysis uses variants of the EBRI-ERF RIPM and RSPM to provide quantitative assessments of the future financial security implications of a move away from defined benefit promises and from annuity payments ("traditional" employer-provided pensions).
  • The analysis provides preliminary results on the impact of benefit accrual freezes for pension plans, modifications to cash balance plans, lump-sum distributions of retirement benefits, and payment of retirement accumulations as life annuities. Major findings include:
  • Freezing benefit accruals for all private defined benefit plans beginning in 2005 would have a larger impact (as measured in terms of average real dollar decrease in first year surplus) on younger cohorts. The average annual decrease for those in the youngest cohort modeled (born between 1961 and 1965, inclusive, or those currently between ages 39 and 43) would be $4,886 for families, $2,752 for single males, and $1,686 for single females.
  • Terminating all cash balance plans without a replacement plan in 2005 has a smaller impact than the benefit accrual freeze, since a much smaller percentage of participants would be affected.
  • The last two sensitivity analyses tested the value of annuitization and therefore needed to simulate a significant number of future life-paths to capture the longevity risk experienced by retirees. The output metric used was the median percentage of additional compensation that must be saved annually until retirement for a 75 percent chance of covering simulated expenses.
  • If one assumes that all defined benefit participants take lump-sum distributions at retirement, the average annual increase in needed savings is 14.9 percent, whereas assuming that all individual accounts are annuitized at retirement has an impact twice as large-but in the opposite direction (a 30.0 percent decrease in needed annual savings).
  • Policy makers should take action sooner rather than later in order to create greater regulatory certainty for plan sponsors. Decisions are needed on the status of cash balance pension plans, permanent funding rules, and interest rates to be used in plan calculations, accounting treatment related to using smoothing versus mark-to-market for investment returns and interest rates, and rules and premiums under Title IV of ERISA and PBGC. Until these kinds of policy decisions are made, further erosion of the defined benefit system can be expected to continue. While the decisions made could either slow or speed that erosion, they would at least create an environment in which individuals could better assess what they are likely to have as retirement assets and income, and plan to continue working or to exit the work force, accordingly. A demographic time bomb is ticking, and the time to act is now.

Individual Retirement Accounts

From the February 2004 Notes article, "IRA and Keogh Assets"

  • Total IRA assets decreased for the third consecutive year in 2002, falling from $2.651 trillion in 1999 to $2.333 trillion in 2002.
  • For those individuals with IRA assets and no Keogh assets, the largest percentage of assets were owned by individuals who had only a regular IRA, at 34.5 percent, with individuals having a rollover IRA only following closely behind at 33.9 percent. The third largest percentage of assets (16.4 percent) was owned by individuals who had both regular and rollover IRAs. The remaining percentage (15.2 percent) was owned by IRA holders with a Roth, an education IRA, or some combination of those with another type of IRA.
  • According to the 2001 SIPP Waves 3 and 6, approximately 21 percent of Americans age 21 or older were found to own an IRA, a Keogh, or both in September 2001 and 2002.

From the May 2004 Notes article, "Retirement Accounts and Wealth, 2001"

  • By the last quarter of 2001, 17.6 percent of all Americans age 21 and over owned an IRA and 18.8 percent of all workers owned an IRA. The average balance was $37,015.
  • By the last quarter of 2001, 25.3 percent of all Americans ages 21-64 owned 401(k) account and 30.4 percent of all workers ages 21-64 owned a 401(k) account. The average balance was $36,244.
  • By the last quarter of 2001, the median household wealth of Americans ages 21 and over was $73,708. The median household wealth of households with a retirement account was $171,225 compared with $41,117 for those households without a retirement account.

Elderly and Debt

From the April 2004 Notes article, "Debt of the Elderly and Near Elderly, 1992-2001"

  • The percentage of American families with a head age 55 or older who have some level of debt was 56.0 percent in 2001, down from 53.8 percent in 1992.
  • While the percentage of families with a head age 55 or older with any debt was essentially unchanged from 1992 to 2001, the average total debt level increased from $27,517 in 1992 to $38,771 in 2001 and the median increased from $13,611 to $23,000.
  • From 1992 to 2001, debt payments were approximately 9 percent of family income, ranging from a low of 8.5 percent in 1995 to a high of 10.0 percent in 1998.
  • Housing debt payments grew in relative magnitude versus nonhousing debt payments after 1992, increasing from 56.5 percent of debt payments in 1992 to 62.5 percent in 2001.
  • Debt as a percentage of total assets was virtually unchanged, at approximately 7.0 percent, from 1992 to 1998 but decreased in 2001 to less than 6.0 percent.