Young workers with high earnings who maintain 401(k) plans throughout their working years and retire at 65 are expected to receive median initial annual retirement benefits of 83-85 percent of what they were earning at the end of their careers, according to a study released today by the Employee Benefit Research Institute (EBRI) and the Investment Company Institute (ICI). The study finds that this "replacement rate" of preretirement income—which also includes expected Social Security benefits under the current system—could exceed 100 percent of preretirement income for those with low earnings.

• This Issue Brief develops a model that projects the proportion of an individual’s preretirement income that might be replaced by 401(k) plan accumulations at retirement, under several different projected scenarios. The 401(k) participant behaviors are derived from an analysis of 2.5 million 401(k) participants drawn from the year-end 2000 database collected by the Employee Benefit Research Institute (EBRI) and the Investment Company Institute (ICI) in their collaborative effort known as the EBRI/ICI Participant-Directed Retirement Plan Data Collection Project.

• The most significant factor affecting projected replacement rates at retirement is having access to a 401(k) plan. Projected replacement rates at retirement are reduced significantly when participants are not offered a 401(k) plan in all portions of their careers.

• Most 401(k) participants tend to have contributions in any given year. Thus, projecting that participants always have contributions (their own and/or employer contributions) every year raises projected replacement rates, but not by much compared with the importance of being offered a plan to begin with.

• The model simulations show that participant activities such as taking loans, taking preretirement withdrawals, or cashing out account balances at job change reduce projected 401(k) accumulations and thus replacement rates at age 65. Because loans are forecast to be paid back to the account in full, their effect on replacement rates at retirement in the model is the smallest.

• Even if equity returns in the future are projected to replicate the worst 50-year segment in the Standard & Poor’s (S&P) 500 history (1929 to 1978), 401(k) accumulations are still projected to replace significant proportions of projected pre-retirement income.

• Another projection scenario forecasts participants experiencing a simulated three-year bear market (negative equity returns) either early in their careers, near the middle of their careers, or at the end of their careers. Forecasts of the effects of bear markets on 401(k) balances show that a bear market in equities is projected to have the largest effect the closer it occurs to age 65 (retirement), even though older participants typically have diversified their portfolios away from equities.

• Similarly, a simulated three-year bull market (positive equity returns) is projected to have a larger positive effect on projected account balances and replacement rates the closer to retirement it occurs.

Jan. November 2002 Feb. EBRI Issue Brief (ISSN 0887-137X) is published monthly at $300 per year or is included as part of a membership subscription by the Employee Benefit Research Institute, 2121 K Street, NW, Suite 600, Washington, DC 20037-1896. Periodicals postage EBRI Mar. Figure 5 Figure 9 rate paid in Washington, DC. POSTMASTER: Send address changes to: Figure 3 EBRI Issue Brief, 2121 K Street, NW, Suite 600, Figure 4 Figure 7 Figure 1 Figure 8 Figure 6 Can 401(k) Accumulations Figure 2 EMPLOYEE Distribution of Replacement Rates From Social Security for P a articip b ants Equity Market Investment Returns’ Effect on Median 401(k) Replacement Rates Washington, DC 20037-1896. Copyright 2002 by Employee Benefit Research Institute. All rights reserved, No. 251. a Median Replacement Rates From Social Security for Participants Turning 65 Median Replacement Rates From 401(k) Accumulations Change in Median Replacement Rates from 401(k) Accumulations Relative to Baseline Model Median Replacement Rates for Participants Turning 65 Between 2035 and 2039, a Total Return on S&P 500 Stock Index, 1939–2001 b Distribution of Replacement Rates From 401(k) Accumulations for Diagram of Basic Elements of the EBRI/ICI Turning 65 in the Year Indicated, by Income Quartile at Age 65 Apr. Among Participants Reaching Age 65 Between 2030 and 2039, by Income Quartile at Age 65 BENEFIT in the Year Indicated, by Income Q 401(k) Accumulations for Participants Turning 65 in uartile at Age 65 Assumptions for Participants Reaching Age 65 Between 2030 and 2039, by Income Quartile at Age 65 for Participants Turning 65 in the Year Indicated, by Income Quartile at Age 65 (three-year annual geometric average) Participants Turning 65 in the Year Indicated, by Income Quartile at 401(k) Accumulation Projection Model (percentage of final five-year average salary) (percentage points) (percentage of final five-year average salary) by Income Q (percentage points) uartile at Age 65 RESEARCH the Year Indicated, by Income Quartile at Age 65: (percentage of final five-year average salary) Table of Age 65: High-Tenure Sample The Employee Benefit Research Institute (EBRI) was founded in 1978. Its mission is 40% May (percentage of final five-year average salary) INSTITUTE ho we are High-Tenure Sample ................................................. 15 ® Generate Significant Income for to contribute to, to encourage, and to enhance the development of sound employee benefit (percentage of final five-year average salary) Income Quartile W 401(k) Account Balance Participants Turning 65 Between 2005 and 2009 Income Quartile 49% Start at 48% 56% Continuous 401(k) Coverage and Equity Returns Based on 1926 to 2001 (Baseline Model) programs and sound public policy through objective research and education. EBRI is the only private, nonprofit, Figure 7, Change in Median Replacement Rates from Asset Allocation 45% Participants Turning 65 Between 2005 and 2009 Year-End +31.2% Jun. Contents 1234 +28.9% 1234 Loan Balance nonpartisan, Washington, DC-based organization committed exclusively to public policy research and education on 45% Participants Turning 65 Between 2005 and 2009 43% 401(k) Accumulations Relative to Baseline Model 2000: 59% 52% 100% 38% Annual Income economic security and employee benefit issues. EBRI’s membership includes a cross-section of pension funds, 35% 35% 46% 34% a Assumptions for Participants Reaching Age 65 Bear Market in Equities Occurs: 41% 41% Jul. 29% 33% 29% 89% Assuming Always Have Contributions to 401(k) Plan Account 38% 9.1 103%8.9 6.5 4.6 35 Future Retirees? elects a set of installment payments rather than an See the appendix (see endnote 12) for more discussion accrue to the majority of workers (see also Yakoboski $592 and through $3,567 plus 15 percent of their AIME businesses, trade associations, labor unions, health care providers and insurers, government organizations, and 33% consulting firms. Records were encrypted to conceal the and Wise (October 1995); and, Yakoboski (February tion activity, see Holden and VanDerhei (October 2001). benefit plans or non-401(k)-type defined contribution 24% 23% 29% Overview and Summary Start of career..................................................80% 3 –3.2 –2.9 –3.0 –3.7 27% 28% Assuming Loans Are Never Taken From 401(k) Plan Account Between 2030 and 2039, by Income Quartile at 0.4 0.3 20% 0.4 0.3 77% service firms. 16% b 25% 85% annuity, the amount he or she may reasonably withdraw of these investment categories. (May 1998)). over $3,567 (see the Social Security Administration’s identity of employers and employees. Data provided for 1994). In the model projections, it is assumed that the EGTRRA plans. 86% Middle of career –7.5 –8.0 –8.8 –10.4 Assuming Preretirement Withdrawals Are Never Taken From 401(k) Plan Account 6.7 6.0 6.0 3.8 83% Aug. Effect of Participant Behaviors on Projected Age 65 ........................................................................ Contributions? 16 21% 20% End of career 20% –13.4 –14.1 –15.6 –17.7 59% Assuming Do Not Always Have 401(k) Plan Coverage –27.5 –30.8 –34.7 –39.4 each year after the first year may increase or decrease as by Sarah Holden, ICI, and Jack V Web site, www.ssa.gov, for benefit formulas). anDer each participant include participant date of birth, from provisions do not sunset. hei, Temple University and 69% Amount Contributed? 36 18 44 59 17% 52% Replacement Rates at Retirement ......................... 4 Figure 8, Total Return on S&P 500 Stock Index, 15% Historically (and in the baseline case of the model), See endnote 8. The EBRI/ICI 401(k) Accumulation Projection Model See endnote 43. Assuming Never Cash Out Balance at Job Change 61% 13.3 9.1 6.8 4.7 46% 56% future market conditions affect the account balance EBRI’s work advances knowledge and understanding of employee benefits and their which an age cohort is assigned; participant date of hire, c Sep. 41% 51 Bull Market in Equities Occurs:cd 26 Effect of Investment Returns on Projected 51% Assuming Preretirement Withdrawals Are Never Taken From IRA Balances 11.1 12.8 14.8 18.4 about two-thirds of the time, equity returns in any given Other research has also noted that there is a range of 48% Full Sample does not include contributory traditional IRAs or Roth 1939–2001 See Holden and VanDerhei (November 2001). High-Tenure Sample .................................................................. 62 18 hat we do EBRI Fellow Quartile 1 Quartile 2 Quartile 3 Quartile 4 401(k) plan accounts (and Social Security) is projected to Bernheim, Douglas B., Jonathan Skinner, and Steven Skinner, Jonathan. “Hearing on Retirement Security and assumptions, having entered the model employed at a Income— preretirement consumption activity in retirement. 19 Effect of Cash-Out at Job Change— Asset Allocation there is much less variation by birth cohort. the Treasury, Internal Revenue Service, 2001. Avail- Research Institute, 1999. In order to forecast participant activity inside a —Among EBRI/ICI 401(k) plan partici- Research on partici- ________. “Employee Benefits in Medium and Large account pension assets, especially among later cohorts long enough to vest 60 rate” measures the ability of retirement savings, includ- Moore, James F., and Olivia S. Mitchell. “Projected in benefits. In addition, employ- Figure 4 While Social W limits change over time in the model as legislated in the scfindex.html Accessed 10/23/2002. historical returns. However, if an average participant in that does not lobby or take positions on legislative projected to have Social Security replace 38 percent or 33% importance to the nation’s economy among policymakers, the news media and the public. The definition of preretirement income used by Effect of Investment Retur Although a participant with contributions of 9 percent 31% ns on going forward. Start of career from which a tenure range is assigned; outstanding loan 3.9 3.6 3.9 4.5 26% Replacement Rates at Retirement ......................... 4 year are between –7 percent and 33 percent. Ibbotson individual experiences at retirement. For example, Venti 30 Figure 9, Equity Market Investment Returns’ Effect IRAs. The model only permits participants to have 23% b Loans? 401(k) plan, future incomes must be estimated for each firm that offers a 401(k) plan, the EBRI/ICI 401(k) replace significant proportions of projected preretirement ________. “Lump-Sum Distributions Total $87.2 Billion pant behavior at job change has found that participants pants, asset allocation appears to vary with age. Weinberg. “What Accounts for the Variation in Defined Benefit Pension Plans.” able at: www.irs.gov/pub/irs-pdf/p575.pdf Accessed: In order to estimate replacement rates at Even though the projection horizon is short for It does this by conducting and publishing policy research, analysis, and special reports on employee benefits issues; Testimony Before the born between 1936 and 1964. In addition, Samwick and 27 Private Establishments, 1997.” his or her 20s with the average asset allocation for that proposals. ment risk occurs if the participant changes jobs and ing 401(k) investments, to generate income in retirement Retirement Wealth and Savings Adequacy in the Bulletin compares median 2517 Security replacement Middle of career Economic Growth and Tax Relief Reconciliation Act more of their preretirement income, half of them are 10.2 10.7 11.7 13.8 Memo: 28% Fidelity Investments. Building Futures, Volume II: Oct. different researchers can vary. Results presented of salary and a salary of $40,000 a year will have an In order for the loan not to be treated as a distribu- balance; withdrawals; funds in a participant’s invest- Modeling Growth 9 Projected Replacement Rates at Amount Borrowed? Modeling Growth in 401(k) Plan Participant 21% b What Are Future The EBRI/ICI model primarily is based on 401(k) and Wise (February 2000), using HRS data, conclude End of career on Median 401(k) Replacement Rates Among 25 16.0 16.8 18.6 21.0 6 (2002) data were used (see endnote 11). holding educational briefings for EBRI members, congressional and federal agency staff, and the news media; and rollover IRAs, which serve as receptacles for 401(k) Median Replacement Rates for Typical 401(k) Participant • This Issue Brief develops a model that projects the proportion of an individual’s 50.7 54.0 59.5 67.2 rates tend to fall as income for future retirees (Figure 1, top panel). Alterna- participant. Contributions to the 401(k) plan by partici- participants are assumed to continue to work at employ- retirement, several items must be tracked for each Younger participants tend to have higher percentages of with smaller balances are more likely to cash out those Subcommittee on Oversight of the House Committee on Retirement Wealth Among U.S. Households?” some of these older participants, participants who had 10/23/2002. in 1995.” Over EBRI Notes view and , no. 11 (Employee Benefit Ameri- (EGTRRA) of 2001. Skinner (October 2001) conclude that 401(k) plans Opportunities and Challenges for Workplace Savings (September 1999). age group were to draw the average return for that benefits left at early jobs are not indexed for inflation. and thus to maintain retirees’ standard of living. projected to have replacement rates of 48 percent or Health and Retirement Study.” replacement rates for NBER Working Paper In . 15% tion from the plan, it generally must be repaid within 4 throughout this paper compare retirement income to Participants Turning 65 Between 2020 and 2024 ment portfolios; and asset values attributed to those account balance of about $3,600 in the first year, an Participants Turning 65 Between 2020 and 2024 The Investment Company Institute is the national Accounts....................................................................... sponsoring public opinion surveys on employee benefit issues. 6 EBRI’s Education and Research Fund (EBRI-ERF) participant behavior observed in the EBRI/ICI Partici- that the bulk of the dispersion in wealth at retirement 42 61% Participants Reaching Age 65 Between 2030 and 2005 to 2009 2020 to 2024 Retir balances from previous jobs. Whether the account ement 2035 to 2039 Nov. preretirement income that might be replaced by 401(k) plan accumulations at 37 their account balances invested in equity securities, At job change, participants this study, the replacement rate is calculated for each three different birth income rises, 401(k) tive scenarios designed to analyze the effect of different pants and/or their employers are determined as a ers that offer a 401(k) plan. ________. balances at job change, while larger account balances are 401(k) participant over time. Each participant’s income, Ways and Means can Economic Review already experienced a long working career (relative to Research Institute, October 1999): 4–7. Publication 590 Individual Retirement Ar- . Washington, DC: June 20, 2002. . Vol. 91, no. 4 (September 2001): U.S. Department of Labor. Pension and Welfare Benefits provide pension benefits at retirement at least as well as in America, a Report on Corporate Defined Contribu- portfolio, he or she would be projected to have a nominal Employer risk results from the possibility that the more, and one-quarter of them are projected to experi- No. 6240. Cambridge, MA: National Bureau of Eco- The total return used for bonds, GICs, money market Memo: five years. See Internal Revenue Service (IRS), 75% Publica- five-year average salary immediately before retirement. funds. An account balance for each participant is the otherwise similar participant with salary of $80,000 will 0% performs the charitable, educational, and scientific functions of the Institute. EBRI-ERF is a tax-exempt organization association of the American investment company indus- Source: Tabulations from EBRI/ICI 401(k) Accumulation Projection Model. 65% in 401(k) Plan EBRI/ICI 401(k) Accumulation Projection Model 49% ..... 6 pant-Directed Retirement Plan Data Collection Project. results from the choice of some families to save while 59 d balances are held in the 401(k) or a rollover IRA is not 2039, by Income Quartile at Age 65 3 7 ......................... 19 retirement, under several different projected scenarios. The 401(k) participant behav- 60% Retirees Projected to decide whether to leave their 401(k) balance at the Smith, James P. “The Changing Economic Circum- participant activities and varying investment return percentage of the participant’s income. In addition, the ________. “Debunking the Retirement Policy Myth: while older participants tend to favor fixed-income more likely to roll over into an IRA. contributions, loans, withdrawals, asset allocations, and 832–857. their age) at their year-end 2000 employer were used as rangements (IRAs) Median Replacement Rates for Typical 401(k) Participant . Washington, DC: U.S. Department To the extent that or, in most cases, better than defined benefit plans. Administration. return of about 9- employer may change the prospective terms of the plan individual at age 65. nomic Research, October 1997. Private Pension Plan Bulletin, /4 50.7 percent in that year in the baseline The replacement rate compares 401(k) W 54.0 ithdrawal? cohorts between the 59.5 67.2 accumulation replace- EBRI Year Participant TLoans— urns 65 58% ence replacement rates of 63 percent or more (Figure 5, Most 401(k) plan participants are in plans that 58% a tion Plans. Boston, MA: Fidelity Investments, 2001. funds, and other investments was based on Ibbotson’s However, results relative to final working-year salary Quartile 1 Quartile 2 tion 575 have twice as much. However, the saving rate and ratio Quartile 3 . (This IRC exception also applies to loans that Quartile 4 Summar supported by contributions and grants. y The American Savings Education Council (ASEC) and the Consumer Health sum of the participant’s assets in all funds. Change in median replacement rate for 401(k) accumulations relative to final five-year average salary. This is the first-order 41% 53% difference and does not take into account –1.0% • try. Its membership includes 8,982 open-end investment Even if equity returns in the future are projected to 40%aa Dec. Participant Activity Inside the 401(k) Plan............... 7 In addition, information taken from other surveys is other similarly situated families (in terms of lifetime 46% Amount Withdrawn? Social Security 401(k) Accumulation significant. Asset allocation and investment returns are Social Security and 401(k) Accumulation iors in the model are based on the year-end 2000 database collected by the Employee 43% ment rates tend to rise previous employer, cash it out, or roll it over into an IRA. Burman, Leonard E., Norma B. Coe, and William G. income projections are used to determine final salaries situations are also examined. The key finding is that the investment returns are projected from year-end 2000 securities, such as guaranteed investment contracts lower income participants tend to have lower account stances of the Elderly: Income, Wealth, and Social an approximation to better estimate full career behavior changes in participant behavior that might occur as a result of changing the activity in question. of the Treasury, Internal Revenue Service, 2001. Lifetime Jobs Never Existed for Most Workers.” 33% EBRI allow loans; however, most participants do not borrow Furthermore, Poterba, Venti, and Wise (November 2001) Abstract of 1998, Form 5500 Annual Reports or may enter bankruptcy (in which case, insured pension Munnell, Alicia H., Annika Sundén, and Catherine the individual’s initial annual retirement income gener- case. Similarly, if an average participant in his or her bottom panel). The wider variation within the lower full sample and the . No. 11 Education Council (CHEC) are programs of EBRI-ERF. They are coalitions of private- and public-sector institutions Gustman, Alan L., and Thomas L. Steinmeier. “What are used to buy a main home, regardless of their repay- long-term government bonds total returns from the were substantially similar. For a discussion of the of account balance to salary are identical between those 30% replicate the worst 50-year segment in the Standard 28% 14 companies (“mutual funds”), 513 closed-end investment Participant Behavior Over Working Career b .............. 9 EMPLOYEE Quartile 1 Quartile 2 Quartile 3 25% Quartile 4 used to model participant income changes over time; earnings) chose to spend. They conclude that very little Source: Tabulations from the EBRI/ICI 401(k) Accumulation Projection Model. 60 NOTE: Available public data sets on households include the An extensive appendix detailing the elements of 26 The ratio of the income generated in the first year of retirement from 401(k) accumlations to final five-year average salary (p modeled in the same way in both the 401(k) and rollover ercentage) for the baseline model. Participant Accounts Benefit Research Institute (EBRI) and the Investment Company Institute (ICI) in if they chose never to cash those balances with the goals of public education on saving, retirement planning, health insurance, and health quality. with income. For most important factor affecting projected retirement and measure replacement rates. The income forecast for (GICs) and bond funds. Each year in the model, assets until the participant retires. For simplicity, it is assumed balances, Gale. “Lump Sum Distributions from Pension Plans: Security.” of these birth cohorts. Among the high-tenure sample Issue Brief Available at: www.irs.gov/pub/irs-pdf/p590.pdf Ac- Maxwell School of Citizenship and Public no. 197 (Employee Benefit Research from their accounts. forecast that the average 401(k) balances of people who 21% Receive Fr People Don’t Know About Their Pensions and Social (Winter 2001–2002). amounts may be less than those promised). However, in ated by projected 401(k) accumulations with his or her 60s with the average asset allocation of that age group income quartile primarily results from the higher Taylor. “What Determines 401(k) Participation and At year-end 2000, EBRI/ICI data om Social high-tenure sub- 21% beginning of 1926 to the end of 2001 (percent per annum ment term.) Furthermore, unpublished ICI data from a 11 impact of the definition of preretirement income used on a –7.4% two participants. Research on the relationship of account 17% 2002 companies, and six sponsors of unit investment trusts. & Poor’s (S&P) 500 history (1929 to 1978), 401(k) The bear market consisted of three consecutive years of –9.3 percent annual returns, which reduced equity holdings (equity funds, company stock, and the equity 15% What Are Future Retirees Projected to Receive From frequency of and activities associated with job change; of the dispersion was explained by chance events or asset Participants Turning 65 Betw the model analyzed in this report is available online at IRA accounts. The only differences between the two Federal Reserve Board’s Survey of Consumer Finances een 2020 and 2024 –9.3% their collaborative effort known as the EBRI/ICI Participant-Directed Retirement Leave Balance, Cash Out, or Roll Over income is having access to a 401(k) plan. The study also each participant is based on regression results analyzing are rebalanced based on these changing patterns as that all individuals retire at age 65. As participants age, out at job change, they would see an increase in their The 401(k) plan is only about 20 years old, which means Recent Evidence and Issues for Policy Research.” Affairs, Center for Policy Research, participants, on average, 401(k) accumulations are Institute, May 1998). cessed: 10/23/2002. Not-Al Policy Brief ?—Studies of ways-401(k) Coverage and Equity Returns Based on 1926 to 2001 . No. 8/ U.S. General Accounting Office. “401(k) Pension Plans: will reach retirement age in 2035 are projected to be defined benefit plans, employees generally do not projected five-year average salary immediately before were to experience an average year, he or she would be Contributions?” CRR Working Paper. sample of participants 2000-12. Chest- example, among BENEFIT indicate only 12 percent of participants analyzed had a marginal benefit per additional dollar earned at lower Security: An Analysis Using Linked Data from the Asset Allocation compounded annually; see Ibbotson (2002)). replacement rates, see Steuerle, Spiro, and Carasso portion of balanced funds) by 25 percent by the end of the three-year period. 401(k) household survey (see ICI (Spring 2000) for the balance to salary has found that there is little variation Source: Tabulations from EBRI/ICI 401(k) Accumulation Projection Model.Its mutual fund members manage assets of approxi- accumulations are still projected to replace significant 111% { Social Security and 401(k) Plan Accumulations and IRA activities. The EBRI/ICI model focuses on allocation choices. In addition, see Bernheim, Skinner, b the ICI Web site: www.ici.org (SCF), the University of Michigan’s Health and Retire- 8 52 63 1 vehicles, as modeled here, are (1) loans are not permitted Plan Data Collection Project. 104% Rebalance Portfolio that individuals retiring today could have participated in Middle of career is defined as the years the participant is 39, 40, and 41 years old. participants reaching distribution of 401(k) balances at job change find that finds that even under situations of relatively brief the income paths by age of individuals in the Current 1997. Syracuse, NY: Syracuse University, 1997. ________. “Large Plan Lump-Sums: Rollovers and Investment Company Institute. participants age, while still preserving an individual they may change jobs, and with job change, they may roll replacement rates at retirement. Among participants National Tax Journal projected to generate income to replace at least 41 per- . Vol. LII, no. 3 (September Quarterly Supplemental loan outstanding. Among those participants with an roughly the same size as the present value of their Social Loan Provisions Enhance Participation But May directly bear investment risk. retirement. projected to have a nominal return of about 8- income levels. nut Hill, MA: Center for Retirement Research at who had long tenure /2 percent Note: The model assumes participants have continuous full working careers. EBRI Issue Briefs Health and Retirement Study.” are monthly periodicals providing expert evaluations of NBER Working Paper. published survey results) suggest that the vast majority (May 2000). Security and 401(k) in that ratio across salary groups (see Holden and 75% RESEARCH 98% 38 proportions of projected preretirement income (Figure c mately $6.4 trillion, accounting for approximately at Retirement? ........................................................... ur publications 10 Whereas current retirees cannot reflect the experience of 401(k) plan participants and is distinct from the EBRI- With an inflation rate of 3.3 percent per year in the and Weinberg (September 2001). The bull market consisted of three consecutive years of +31.2 percent annual returns, which increased equity holdings (equity f 24 c ment Study (HRS), the U.S. Census Bureau’s Survey of 38 unds, company stock, and the equity O from IRAs, and (2) withdrawal rules and behaviors vary d –20% participant’s asset allocation preferences relative to the turning 65 between 2030 and 2039, those in the lowest 401(k) plans for at most half of a typical full working Both the level and employee benefit issues and trends, as well as critical analyses of employee 89% in the baseline case. In 401(k) plans, participants face employment Because the future is uncertain, the EBRI/ICI for their age group at Starr-McCluer, Martha and Annika Sundén. “Workers’ age 65 between 2035 larger account balances tend to be rolled over and periods (e.g., three years) of negative equity market Population Survey (CPS) data. account balances into IRAs or cash out the balance 1999): 553–562. cent of projected preretirement income for those Cashouts.” Data. Washington, DC: Investment Company Insti- EBRI Issue Brief no. 188 (Employee Full Sample outstanding loan balance, the loan represented, on Effect of Preretirement 401(k) Withdrawals— Security benefits. On the other hand, Wolff (2002) argues No. 7368. Cambridge, MA: National Bureau of Eco- Affect Income Security for Some.” Boston College, December 2000. High-Tenure Sample Among participants reaching age 65 between Letter Report Although . 10/01/ pants are a little less likely to have contributions, their of 401(k) participants who took a loan from their 401(k) Quartile 1 Quartile 2 VanDerhei (November 2001)). Quartile 3 Quartile 4 20 portion of balanced funds) by 126 percent by the end of the three-year period. • The most significant factor affecting projected replacement rates at retirement is 95 percent of total industry assets, and represent more 1, bottom panel). an entire career with exposure to 401(k) plans, projec- forecast, the real return for the average participant Projected Replacement Rates From In addition, to screen for part-time employees or ERF (Education and Research Fund) Retirement Income INSTITUTE 1939 1942 1945 1948 1951 1954 1957 31 1960 43 1963 1966 1969 Income and Program Participation (SIPP), or the U.S. between the two vehicles. 1972 1975 1978 1981 1984 1987 1990 1993 1996 9 1999 benefit policies and proposals. Each issue, ranging in length from 16–28 pages, thoroughly explores one topic. ® 52% EBRI smaller account balances tend to be cashed out. Copeland, Craig. “Lump-Sum Distributions: An Update.” performance, or returns from the worst historical 50-year growth of an individual’s income over time are depen- Knowledge of Their Pension Coverage: A Reevalua- accumulated at a previous employer. All of these ele- average participant in the age group. income quartile at age 65 are forecast to replace career. As more American workers participate in 401(k) participants reaching age 65 between 2005 and 2009 52 Benefit Research Institute, August 1997). tute. d If a that among near-retirees in 1998, only those with wealth 97, GAO-HEHS-98-5. Washington, DC: October 1997. 401(k) Accumulation Projection Model risk, but it is not as great as in defined benefit plans. Poterba, James M., Steven F. Venti, and David A. Wise. Because exposure of 401(k) participants to 75% year-end 2000. When examines and 2039, 401(k) average, 15 percent of the 401(k) account balance 2035 and 2039, one-quarter of those in the highest replacement rates are increased the most. For example, 48% few participants make preretirement withdrawals from nomic Research, September 1999. See endnote 50 for the explanation of the PIA for- The ratio of the income generated in the first year of retirement from 401(k) accumulations to final five-year average salary ( plan repaid the loan in full within five years. percentage) for the baseline model. 46% having access to a 401(k) plan. Projected replacement rates from 401(k) accumula- • 61 than 90 million individual shareholders. Another projection scenario forecasts participants Plan Accumulations Investment Returns 42% tions of what current 401(k) participants might achieve experiencing an average year in his or her 20s is pro- individuals who had worked at their current employer Social Security Notes ...................................................... is a monthly periodical providing current information on a variety of employee benefit topics. 11 Other research has also found small effects, on 65% EBRI’s Washington Projection Model (see VanDerhei and Copeland (April Census Bureau’s and Bureau of Labor Statistics’ Current Year Sarah Holden is senior economist, Research plans and become increasingly responsible for their own 45 accumulations are period for the U.S. equity market throughout the entire participant leaves the account balance with the previous dent on an individual’s education and gender. This tion.” ________. “Retirement Program Lump-Sum Distribu- ________. “401(k) Plan Participants: Characteristics, ments are incorporated in the EBRI/ICI 401(k) 13.3 percentage points more of their preretirement EBRI Notes who are in the lowest income quartile and at least Finance and Economics Discussion Series . No. 7 (Employee Benefit Research . Hewitt Associates, LLC. “Cashing Out Your Future?” U.S. Social Security Administration. Office of Policy. (including the loan balance). Each year, for any partici- their 401(k) accounts, withdrawals are permanently holdings above $1 million saw consistent increases in income quartile at age 65 are projected to have Social Participants in 401(k) plans are always 100 percent equity market risk is an area of policy concern, several several potential scenarios. In each scenario, each “The Transition to Personal Accounts and Increasing participants in the among participants reaching age 65 between 2030 and 58% 58% mula, which shows that an additional dollar of AIME Thus, participants who held company stock in their 28% tions at retirement are reduced significantly when participants are not offered a 28 experiencing a simulated three-year bear market Bulletin provides sponsors with short, timely updates on major federal developments in employee benefits. 52% EBRI’s Projected Replacement Rates From 26% 5 In the model, it is assumed that if the participant at future retirement dates are needed. A standard 2001; July 2002)). jected to be 6-1/2 percent. For the average participant in preretirement income. For example, among individuals for only part of the year in 2000, participants in the Population Survey (CPS). (See the bibliography for Web average (for example, see Engelhardt (June 2001) and benefit calculations in their present form and does not 24% P Particip articipants ants T Turning 65 Betw urning 65 Between 2035 and 2039 een 2035 and 2039 Department, at the Investment Company 22% employer, then the participant moves through the 401(k) information is not available in the EBRI/ICI database. Investment Returns— income had they never cashed out a 401(k) balance, retirement security, the question of whether their 401(k) To project 401(k) participant 21% 21% vested in their own contribution to the plans, but must rate-of-return scenarios were examined in the projection participant’s income, contributions, loans, withdrawals, In this effort, EBRI and ICI have collected data from EBRI/ICI database are projected to generate projection period, projected retirement income from No. 1999-5. Washington, DC: Federal Reserve Board, Accumulation Projection Model. Institute, July 2002): 1–8. 59 percent for those in the highest income quartile tions: Hundreds of Billions in Hidden Pension Contributions, and Account Activity.” ICI Research pant not already in the process of paying down an removed (unlike loans—where the amount is projected to retirement wealth (after inflation) compared with their Press Release. Lincolnshire, IL: Hewitt Associates, Annual Trustees Report Security replace 14 percent or less of their preretirement 45% Retirement Wealth: Macro and Micro Evidence.” . Washington, DC: U.S. Social 2039, those in the lowest income quartile at age 65 are generates higher marginal benefits the lower the AIME. 401(k) plans continue to hold company stock in their Source: Calculation using Standard & Poor's 500 total returns (see endnote 11) from Ibbotson (2002). Fundamentals of Employee Benefit Programs 63% offers a straightforward, basic explanation of employee benefit programs 15% 401(k) plan in all portions of their careers. 40% (negative equity returns) either early in their careers, 37% methodology to generate such projections is building a his or her 60s experiencing an average year, the pro- who were in their late 20s in 2000 and are projected to EBRI/ICI database with an annual salary less than the 401(k) Accumulations 1 ........................................... 11 changes jobs within the five-year repayment window, Poterba, Venti, and Wise (August 1999)). consider the impact of Social Security reform. Thus, sites.) at Retirement? 16 69% 69% some of their members that serve as plan record-keepers Institute (ICI), and Jack VanDerhei, Temple ________. “Asset Allocation: IRAs and 401(k)s.” 401(k) accumulations are forecast to be significant. 10 plan activities in his or her new job, until another job Consequently, based on each participant’s age and January 1999. account balances at retirement, the assets held in the while those in the highest income quartile are projected accumulations (Figure 4, top panel). Among participants reaching age Income.” SeriesThe starting component for a participant in the (Investment Company Institute, Spring 2000). EBRI Issue Brief will produce sufficient income in retire- no. 146 (Employee Benefit EBRI counterparts in 1983. Security Administration, February 2002. vest in the employer contributions. Typically, the vesting model. The baseline case uses the longest historical time and asset allocations are projected every year from year- NBER Working Paper. However, Wolff (2002) fails to No. 8610. Cambridge, MA: analyzed based on income to replace at existing loan, the model assigned a probability to income, half of them are projected to have replacement Change Jobs? projected to replace 9.1 percentage points more of their equities on the projected median replacement rates from in the private and public sectors. The EBRI Databook on Employee Benefits be repaid to the account). If participants never made 30 May 2000. income in retirement. This paper uses a standard is a statistical reference volume on employee A complete analysis of preparedness for retirement IRAs. For a study comparing the asset allocation of 62% 61% 53 near the middle of their careers, or at the end of their 48% 57% model based on typical behaviors observed today to see jected real return is about 5-1/4 percent. turn 65 between 2035 and 2039, the median individual age-equivalent for a 25-year-old earning $5,000 were Variation of Projected Replacement Rates Within The effects presented do not take into account the 56% then the remaining loan balance is immediately repaid Social Security replacement rates do not vary much by 62 and administrators. The EBRI/ICI data collection project University, is research director of the Employee ment is a significant policy concern. This benefit programs and work force related issues. 51% Issue Brief 15 52% Steuerle, Eugene, Christopher Spiro, and Adam Carasso. least 51 percent of change occurs. If the participant then chooses to cash out income, an education level and gender are assigned to Notes Ippolito, Richard A. model is the 401(k) account information for year-end accounts must earn investment returns over the 401(k) accumulations. Although younger participants to replace 4.7 percentage points more, compared with the 65 between 2020 and 2024, half of the lowest income Research Institute, February 1994). . No. 10 (Employee Benefit Research Institute, Pension Plans and Employee Perfor- • Most 401(k) participants tend to have contributions in any given year. Thus, project- Holden, Sarah, and Jack VanDerhei. “401(k) Plan Asset ________. “Formula for Primary Insurance Amount.” whether a given participant would borrow from his or withdrawals from their 401(k) plans, participants project defined contribution plan balances at retirement, rates of 15 percent or less, and three-quarters of them period available for S&P 500 total returns: 1926 to period in 401(k) plan matching employer contributions end 2000 until the participant retires. In addition, in the simulation model methodology to project 401(k) accumu- National Bureau of Economic Research, November birth cohort without preretirement income if they always had contributions Vesting refers to the amount of time a participant would also require estimating retirement income from For research covering the confusion evidenced in household (contributory and rollover) IRA and 401(k)- 38% careers. Forecasts of the effects of bear markets on Birth Cohorts ........................................................ 13 changes in other behaviors that might result from in full to the account. a where individuals end up if they continue down their in the lowest income quartile at age 65 is projected to see also not included in the analysis. 61 33% birth cohort (generation). Half of participants in the a 39 64 Benefit Research Institute (EBRI) Fellows 39 the 401(k) balance, only the 401(k) account balance at each EBRI/ICI 401(k) participant. The model projects Yuh, Yoonkyung, Sherman Hanna, and Catherine October 2000): pp. 5–9. “Measuring Replacement Rates at Retirement,” develops a model that projects the proportion of an 2000 in the EBRI/ICI database. The year-end 2000 participant’s projected working career. In the EBRI/ICI baseline case (Figure 7). quartile at age 65 are projected to replace at least mance: Evidence, Analysis, and Policy Social Security . Chicago: The 401(k) Accumulation whereas he does project the present value of Social Available: www.ssa.gov Accessed 10/23/2002. has been shorter than in defined benefit plans. 2001. model, participants may change jobs, and with a job is the most comprehensive source of 401(k) plan partici- 2001. The first alternative scenario uses the worst Social Security and 401(k) Accumulation regard for tenure (the Em- preretirement income tend to concentrate more of their accounts in equity ing that participants always have contributions (their own and/or employer contribu- her account based on the individual’s age, tenure, and reaching age 65 between 2030 and 2039 are projected to Allocation, Account Balances, and Loan Activity in are projected to experience replacement rates of 18 per- lations at age 65 for a very large sample of current compared with their baseline results when they didn’t Ef defined benefit plans, all IRAs, and possibly other For an explanation of the Ibbotson data used, see fect of Participant Behaviors on 28% household survey responses, see Gustman and type plan balances, see Copeland (October 2000). must work before earning a nonforfeitable right to a 25% 401(k) balances show that a bear market in equities is Leave Balance, Cash Out, current paths. This study builds such a model, the EBRI/ Effect of Participant Behavior on Projected Social Security and 401(k) accumulations replacing changing the behavior in question. There are potentially several sources of income in lowest income quartile at age 65, reaching age 65 (Figure 1, top and bottom panels). Nevertheless, 401(k) experiencing three contiguous years of controlled equity 21 29 pant-level data available to date. The EBRI/ICI data are Program. Special thanks to Craig Copeland, Engelhardt, Gary V. “Pre-Retirement Lump-Sum their most recent (soon-to-be) previous employer is annual incomes for each individual based on his or her Straight Talk on Social Security and Retirement individual’s preretirement income that database, 401(k) plan participants’ account balances are EBRI/ICI database contains records for 11.8 million 58 percent of salary, while half of the highest income Phillips Montalto. “Mean and Pessimistic Projections University of Chicago Press, 1997. might be replaced VanDerhei, Jack, and Craig Copeland. Security and defined benefit plan benefits at retire- change, they may set up rollover individual retirement 50-year time period for the S&P 500, 1929 to 1978, which ployer risk for 401(k) participants occurs if the employer ________. “Pre-Retirement Cashouts and Foregone full sample), projected Kansas Future for half of those in the tions) every year raises projected replacement rates, but not by much compared with salary. Once it is determined that a loan would be taken, cent or less (Figure 5, bottom panel). Similar variations 25% always have contributions (Figure 7). For the median endnote 11. securities, because little of the final balance has been See Holden and VanDerhei (November 2001). replace 3.8 to 6.7 percentage points more of their 2000.” pension benefit (see U.S. DOL (September 1999)). 401(k) plan participants. A number of future scenarios Given that participants age 59-1/2 or older may take ICI Perspective, Vol. 7, no. 5; and 18% EBRI Issue defined contribution plans. For examples of research Contact EBRI Publications, (202) 659-0670; fax publication orders to Steinmeier (September 1999) and Starr-McCluer and 21% or Roll Over? 18% 15% Projected Replacement Rates at 46 projected to have the largest effect the closer it occurs 14% ICI 401(k) Accumulation Projection Model, using recent 103 percent of their preretirement income (Figure 1, top Replacement Rates at Retirement ubscriptions/orders ........................... 14 retirement: (1) Social Security benefits; (2) income from See the appendix (see endnote 12). IRA withdrawal between 2005 and 2009, are projected to replace at least 17 accumulations are still projected to replace substantial market returns either early, mid, or late in their careers. 54 S unique because they cover a wide variety of plan admin- senior research associate at EBRI, who tabu- by 401(k) plan accumulations at retirement, under lowest income quartile cashed out; balances in IRAs or held at any earlier personal characteristics and growth in the national Pension Distributions and Retirement Income Secu- Policy Johnson, Richard W. “The Gender Gap in Pension 401(k) plan participants; of these, 2.5 million are a accumulated early-on, a bull market in equities that is Effect of IRA Withdrawals— identified by type of investment objective. In the model, quartile are projected to replace 75 percent of salary or of Retirement Adequacy.” . No. 24. (Urban Institute, May 30, 2000). Quartile 1the importance of being offered a plan to begin with. In the model, participants Financial Services Review Quartile 2 (202) 775-6312. Subscriptions to . the percentage of the account balance borrowed is preretirement income compared with the baseline ment. Brief Retirement Income Assessment Project within income quartiles and birth cohorts are projected accounts (IRAs) or cash out the balance accumulated at a concentrates on a time period excluding some strong bull changes prospective terms of the plan and, if partici- are examined, but the bottom line is that 401(k) accumu- Retirement Saving: Implications for 401(k) Asset Quartile 3 no. 239 (Investment Company Institute and EBRI Issue Briefs Quartile 4 replacement rates at are included as part of . A Project of the individual in the highest income quartile, 4.6 percentage It is assumed that contribution amounts are not penalty-free withdrawals, two separate regression addressing preparedness for retirement, see Scholz Sundén (January 1999). 40 22 63 48 to age 65 (retirement), even though older participants Effect of Investment Returns on Projected With a projected inflation rate of 3.3 percent, the real This difficulty with tenure only occurs in the initial However, if poor investment returns contribute to the 401(k) participant behavior observed in the EBRI/ICI portions of projected preretirement income, ranging from Retir panel). ement The median individual in the highest income private pension plans, whether defined benefit, defined activity in the model is based on Sabelhaus (December 45 percent of their preretirement income with Social EBRI membership, or as part of a $199 annual subscription to lated Current Population Survey (CPS) and EBRI Notes and EBRI Issue Briefs. Individual copies are Toder, Eric, Cori Uccello, John O’Hare, Mellisa previous employer(s) are not cashed out. If the partici- average wage. rity: Evidence from the Health and Retirement several different projected scenarios. who do not cash out balances at job change may roll rates of returns are projected for three investment representative sample containing information for every more (middle panel). Results for the full sample and Vol. 7 (1998): 175–193. Wealth: Is Women’s Progress in the Labor Market EBRI Education and Research Fund and the Milbank previous employer. markets but including several severe bear markets (the istrators and record-keepers and, therefore, a wide range pants hold company stock, they face the potential Accumulation.” NBER Working Paper retirement appear to . No. 7314. and at least 69 percent simulated to occur early in a participant’s career does calculated using a regression equation, but subject to scenario, depending on income quartile (Figure 7). Employee Benefit Research Institute, November to occur for participants projected to retire in other time lations for workers with continuous 401(k) coverage over points more of income are projected to be replaced if they influenced by the change in frequency. equations are used: one for participants in their 60s and (August 2001); Uccello (July 2001); Engen, Gale, and Continuous 401(k) Coverage and Equity Returns Based on 1929 to 1978 16 typically have diversified their portfolios away from { • The model simulations show that participant activities such as taking loans, taking d Participant-Directed Retirement Plan Data Collection return for the average participant experiencing an quartile at age 65 is projected to replace 85 percent of his selection of the participants because the previous Replacement Rates available with prepayment for $25 each (for printed copies) or for $7.50 (as an e-mailed electronic file) by calling .................................................... c 17 contribution, or both; (3) income from IRAs, whether 2000). Wolff analyzes SCF data; however, a Federal Reserve distress termination of the plan, participants face the Security, while half of the lowest income quartile partici- Rollover IRA? a projected median of 41 percent for the lowest income 20 32 Effect of Three-Year Bear Market—Figure 9 presents the Full Sample of plan sizes offering a variety of investment alterna- Survey of Consumer Finances (SCF) data; Luis High-Tenure Sample pant chooses to roll the balance over into an IRA, then a Study.” Favreault, Caroline Ratcliffe, Karen Smith, Gary categories: diversified equity funds them into IRAs. However, once a participant has a element of the model. high-tenure sample differ only slightly among partici- Equalizing Retirement Benefits?” Aging Studies Program Paper Because current retirees cannot accurately Because the EBRI/ICI database Brief Series and the equity . No. 23. Syra- Participants . No. 1 Turning 65 BetwMemorial Fund. Washington, DC: Employee Benefit 1929 stock market crash, the 1937 crash, and the sell-off negative consequences of inadequate diversification. een 2035 and 2039 Cambridge, MA: National Bureau of Economic Re- In this study, the EBRI/ICI 401(k) Accumulation vary widely by birth for half of those in the IRC regulations. All loans originated over the projection periods as well (Figure 5, top and middle panels, for continuously had contributions compared with the not have a large effect on account balances at retirement. Quartile 1 Quartile 2 one for participants younger than 60. 2001). a full working career are projected to generate substan- Quartile 3 Quartile 4 Uccello (May 2001); Montalto (April 2000); the Social EBRI/ICI 401(k) Accumulation © 2002. 55 • The most significant factor affecting projected replace- 13 equities. Asset Allocation, Investment 89% EBRI or from www.ebri.org. Change of Address: EBRI, 2121 K Street, NW, Suite 600, Washington, DC 20037, (202) average year in his or her 20s is a projected 4 percent in accumulations left at other employers or in rollover IRAs Other research has found that savings rates greatly preretirement withdrawals, or cashing out account balances at job change reduce risk of receiving the insured pension amounts (which Project Conclusion or her preretirement income with Social Security and to project what 401(k) participants might ...................................................................... 19 contributory, rollover, or both; (4) income from other Board researcher, Kennickell (1998), has criticized pants reaching age 65 between 2020 and 2024, are 33 44 92% 27, 28 quartile at age 65 to a projected median of 56 percent for 55 projected effect of the timing of a bear market in equities 47 tives. Alonso, research associate at EBRI, who main- reflect the typical experience of individuals working an highest income rollover IRA is created for the participant. Contributions— cuse, NY: Maxwell School of Citizenship and Public Burtless, and Barry Bosworth. rollover IRA established, preretirement withdrawals The median replacement rates at age 65 for participants covers the 401(k) account balance at the participant’s portion of balanced funds, pants reaching 65 between 2035 and 2039 because they (Urban Institute, March 1999). For each year in the model, it is deter- company stock (the Modeling Income in the In this last ________. “Contribution Behavior of 401(k) Plan Partici- are paid down over the subsequent five years Effect of Lapses in 401(k) Coverage— Research Institute, July 16, 2002. example). Projection Model is used first to project replacement from 1973 to 1974). In this scenario, an average partici- tial income at age 65. The most significant factor in search, August 1999. Participants in 401(k) plans directly bear cohort. For example, The baseline results and baseline model. See endnote 8. 80% Security Administration’s Modeling Income in the Near Employee Returns, IRA Withdrawals? 659-0670; fax number, (202) 775-6312; e-mail: Publications Subscriptions@ebri.org. 30 Membership Information: Inquir- ment rates from 401 (k) accumulations at retirement 49 • Similarly, a simulated three-year bull market (posi- Projection Model the lower-equity-return scenario. For the average Bibliography are not available on the EBRI/ICI database. Over the influence the distribution of wealth at retirement (for ................................................................... projected 401(k) accumulations and thus replacement rates at age 65. Because loans 19 75% may be lower for some participants). For the most recent analysis, see Holden and Wolff’s previous SCF-based research. expect from their 401(k) accumulations at retirement the highest income quartile. 401(k) accumulations in the first year of retirement. for participants reaching age 65 between 2030 and 2039 individual savings (in some cases, including home projected to replace at least 49 percent (Figure 3). tains the EBRI/ICI project databases; and Quartile 1 Quartile 2 Quartile 3 Quartile 4 case, the participant is projected to move through 401(k) mined whether a participant’s account will receive a Kennickell, Arthur B. “Comments on ‘Recent Trends in Affairs, Center for Policy Research, Syracuse Univer- Near Term—Projections of Retirement Income entire career with the availability of 401(k) plans, this employer’s stock), and all other investments (bond funds, from IRAs are also possible in the projection. Among current employer and does not include rollover IRAs or are young in the year-end 2000 database and essentially 71% ________. “The Changing Face of Private Retirement rates for participants in a “baseline case.” The baseline pant in his or her 20s with the average portfolio investment risk. Whereas the baseline results of the ________. “Lump-Sum Distributions from Retirement 75% among those partici- quartile (Figures experiencing a bull market early in their careers are earn the bond rate of return. presented above assume that EBRI/ICI 401(k) partici- pants.” reducing projected replacement rates from 401(k) ICI Perspective, Vol. 7, No. 4; and EBRI Issue Benefit 25th Percentile Median 675th Percentile 69% 72% 72% Term (MINT) projections summarized in Toder, Uccello, ies regarding EBRI membership, and/or contributions to EBRI-ERF should be directed to EBRI President Dallas 48 is having access to a 401(k) plan. Projected replace- tive equity returns) is projected to have a larger Because retirees may have lower living expenses (no example, see Samwick and Skinner (October 2001), Venti are forecast to be paid back to the account in full, their effect on replacement rates at VanDerhei (November 2001). under several scenarios. participant in his or her 60s experiencing an average Endnotes projection, the model tracks all 401(k) accumulations— ........................................................................ 23 equity); and (5) income from continued employment, To construct income quartiles, income cutoffs at age Similarly, half of the participants in the highest income 34 compared with their baseline results. If a market down- 62% 64 17 Darrin Helsel, research analyst at ICI, who plan activities in his or her new job, but also manages contribution, whether from the participant, the em- sity, June 2001. Through 2020 for the 1931-60 Birth Cohorts study reports the results of a model built to project what the bond portion of balanced funds, participants reaching age 65 between 2030 and 2039, 401(k) account balances left at previous employers’ have a full career before them in the model. the Size Distribution of Household Wealth,’ by Ed- Salisbury at the above address, (202) 659-0670; e-mail: salisbury@ebri.org GICs, money . Washing- Plans.” case assumes continuous employment, continuous 401(k) experiencing an average year, would have a projected projection model analysis randomly expose participants Savings Plans: Receipt and Utilization.” EBRI Issue Brief no. 232. (Employee Benefit pants close to NBER 1 and 4). Research 401(k) Accumulation Replacement Rates—There is also Effect of Loans— projected to increase by 3.6 to 4.5 percentage points Although most participants are permit- pants always choose an employer that offers a 401(k) This section summarizes the components of the EBRI/ Wolff (2002) uses defined contribution plan balances Brief accumulations for future retirees is not having access to Prior to EGTRRA, vesting practices were subject to , no. 238 (Investment Company Institute, October O’Hare, Favreault, Ratcliffe, Smith, Burtless, and57% 56% children in school, no transportation expenses to and ment rates at retirement are reduced significantly positive effect on projected account balances and 54% Source: Tabulations from EBRI/ICI 401(k) Accumulation Projection Model. Source: EBRI/ICI 401(k) Accumulation Projection Model. year, the projected real return is about 3-1/2 percent. at the current employer when the person reaches age 65, and Wise (February 2000), and Even and Macpherson retirement in the model is the smallest. 65 were determined for each five-year birth cohort that Although projections are always surrounded by 52% perhaps at a part-time job. This report focuses on income quartile at age 65 reaching age 65 between 2005 and 50% 35 1 Modeling Concentrated Equity Market Extremes 48% —To turn occurs early in a participant’s career, little of the 31 Projected Replacement Rates Fr Source: Tabulations from the EBRI/ICI 401(k) Accumulation Projection Model. om Social provided research support. This report is being Institute- 401(k) participants might expect from their 401(k) a 45% Engen, Eric M., William G. Gale, and Cori Uccello. “Are the IRA. ployer, or both. A two-step process is modeled: (1) Will ton, DC: The Urban Institute, September 1999. (depending on income quartile) compared with their those in the lowest income quartile are projected to funds, other stable value funds, other and unknown). plans, account balances of 401(k) plan participants at ward N. Wolff,” Participants Journal of Economic Perspectives, 401(k) Withdrawals plan as they change jobs over the remainder of their 2001, and Employee Benefit Research Institute, Research Institute, April 2001). variation in projected 401(k) accumulation replacement plan coverage, and historical experiences with invest- nominal return of about 7- the same legal constraints for defined contribution and to a distribution of historical rates of return (based on a 401(k) plan in all portions of the individual’s career. For example, if a participant in his or her 20s holds a 48% Working Paper—In some circumstances, partici- . No. 5298. Cambridge, MA: National /4 percent. Similarly, in this retirement (reaching ted to borrow from their 401(k) accounts, most do not ICI 401(k) Accumulation Projection Model used to at the time of the 1998 SCF and defines participants Bosworth (September 1999); Yuh, Hanna, and Montalto Baseline model assumes participants have continuous coverage in 401(k) plans. 41% from work, and possibly no mortgage payments), they when participants are not offered a 401(k) plan in all 43% replacement rates the closer to retirement it occurs. Note: The model assumes participants have continuous full working careers. at all previous employers between year-end 2000 and (March 1998)). divided the cohort into four equal groups of participants b uncertainty, the EBRI/ICI model offers valuable insight model the effect of extreme equity market results on final balance has been accumulated and there is a long future retirees are projected to receive from Social 2009 are projected to replace 20 percent of their Figur 41 es published simultaneously by the Investment Education The 401(k) accumulation includes 401(k) balances at employer(s) and rollover IRA balances. Uccello, Cori E. “Are Americans Saving Enough for there be a contribution? and (2) How much will be Households Saving Adequately for Retirement? A accumulations at retirement after a full working career Historic returns for these three investment categories replace 11.1 percentage points more of preretirement their current employer tend to vary not only with age Vol. 12, No. 3 (Summer 1998. Washington, DC: 37% VanDerhei, Jack, Russell Galer, Carol Quick, and John ment returns—based on an individual’s asset allocations S&P 500 total returns from the beginning of 1926 to the lower-equity-return scenario, an average participant in Bureau of Economic Research, October 1995. age 65 between 2005 reaching age 65 Each participant’s income path is independent of the baseline experience. A bull market in equities is pro- pants are permitted to take withdrawals from their working careers in the projection model. However, many higher percentage of his or her account in equity funds October 2001). rates within birth cohorts. For example, among partici- defined benefit plans, while practically, defined contribu- take advantage of the loan option. In addition, amounts Editorial Board: • Even if equity returns in the future are projected to replicate the worst 50-year Dallas L. Salisbury, publisher; Steve Blakely, managing editor; Cindy O’Connor, production and distribution. Any forecast 401(k) asset accumulations at retirement for near retirement to include persons age 47 and older. (1998); Smith (1997); or Moore and Mitchell (October Security Variation of Pr 28%ojected Replacement Rates c portions of their careers (Figure 1, middle panel). may not need to replace 100 percent of their Full sample contains 2.5 million participants drawn from the year-end 2000 EBRI/ICI database. into the possible future situations of 401(k) participants when the person reaches 65, and all amounts rolled into Endnotes 21% Security and from their 401(k) plan accumulations. By (quartiles). preretirement income with Social Security, while half of participants, the worst and best contiguous three-year 21 time horizon over which to recover. The median replace- and Research 56 57 views expressed in this publication and those of the authors should not be ascribed to the officers, trustees, members, or othe Company Institute as an ICI Perspective. Any r IRA Asset Allocation and Investment Returns— contributed? Analysis of contribution activity among Progress Report on Three Projects.” Retirement?” with exposure to 401(k) plans. A standard methodology are used to create the range of returns possible in any income if they had not taken withdrawals from their but, more importantly, with tenure. Federal Reserve Board Web Site, 1998. Available: Issue in Brief. No. 7, Chestnut Hill, MA: Working Paper. Participants with At Rea. “401(k) Plan Asset Allocation, Account Balances, over the forecast and the range of rates of return histori- end of 2001), this section first examines the effect of an Sabelhaus, John. “Modeling IRA Accumulation and his or her 60s with the average portfolio experiencing an and 2009), half of the between 2035 and Other research has shown that participants in plans d 401(k) plans. However, given the restrictions and than the average participant in their 20s, then that pants turning 65 between 2035 and 2039, three-quarters borrowed typically are paid back into the account. number of times the individual changes jobs and is jected to have a larger effect the closer it occurs to segment in the Standard & Poor’s (S&P) 500 history (1929 to 1978), 401(k) accumu- 15% Holden, Sarah, Jack VanDerhei, and Carol Quick. each participant in a sample drawn from the year-end employers do not offer a 401(k) plan. However, an individual who is 47 years old would have tion plans typically had shorter vesting periods (see U.S. When the EBRI/ 1997). High-tenure sample contains nearly 1 million participants with high tenure for their age at year- preretirement income. Thus, replacement rates may • Figure 1, Median Replacement Rates for Participants Most 401(k) participants tend to have contributions in Within Birth Cohorts The next section of this report discusses key IRAs. sponsors of the Employee Benefit Research Institute, the EBRI Education and Research Fund, or their staffs. Nothing herein is to at retirement. First, several different scenarios are the time the 401(k) participants are projected to reach those in the highest income quartile at age 65 reaching Fund. average total returns on the S&P 500 after the Great Based on the lifetime earnings history generated for each ment rates from 401(k) accumulations at age 65 for 49 views expressed in this report are those of the used in building such a model starts with typical indi- end 2000. 2039 are young at rollover, the asset allocation of the 401(k) balance is EBRI/ICI participants in calendar year 2000 suggests modeled as explained in the appendix (see endnote 12). Prepared for the Third Annual Conference of the Boston College, Center for Retirement Research, lower tenures tend to have lower account balances and given year in the projection. retirement. A simulated three-year bull market increas- IRAs, while those in the highest income quartile are with a loan option contribute higher percentages of www.federalreserve.gov/pubs/oss/oss2/papers/ penalties involved, very few 401(k) participants take assumptions for 401(k) behavior related to contributions, ICI participants experience careers with lapses in 401(k) These are baseline results. As discussed, replacement participant will hold a higher percentage of his or her “401(k) Plan Asset Allocation, Account Balances, and and Loan Activity,” of those in the lowest income quartile at age 65 are cally observed in the United States. To highlight results DOL (September 1999)). However, EGTRRA legally entire projection period that replicates the equity average year, would have a projected nominal return of Withdrawals.” National Tax Journal ICI Perspective lowest income quartile , Vol. 5, no. 1; and . Vol. 53, no. 4, However, if participants were assumed to have never lations are still projected to replace significant proportions of projected pre-retirement 2000 EBRI/ICI database. Using salary, contribution, another 18 years to work, assuming a retirement age of be construed as an attempt to aid or hinder the adoption of any pending legislation, regulation, or interpretative rule, or as understate maintenance of preretirement standards of legal, 11 any given year. Thus, projecting that participants highlights in the structure of the EBRI/ICI 401(k) Turning 65 Between 2035 and 2039, by Income The baseline projection scenario assumed that future All rights examined with the model. In addition, the model is based Depression era were used for the bear and bull market 45 participants experiencing a bear market early in their age 65, some may have rolled some of their 401(k) age 65 between 2020 and 2024 are projected to replace 3 40 23 EBRI/ICI participant, the projected Social Security authors and should not be ascribed to the maintained within the new rollover IRA balance. that approximately 91 percent of participants had Retirement Research Consortium, “Making Hard July 2001. vidual behaviors observed today to forecast where projected to replace 18.4 percentage points more, com- jep.wolff.3.pdf Accessed 10/23/2002. EBRI Issue Brief that are representative of what a full career with expo- investment experience of the worst 50 years of S&P 500 Part 1 (December 2000): 865–875. /4 percent. no. 205 (Investment Company In other model scenarios, the at age 65 are projected year-end 2000 and the many such participants, especially older ones, may have ing equity holdings by 126 percent immediately before salary. For example, see Holden and VanDerhei (October The detailed assumptions governing the evolution of For equity investments in the baseline projec- withdrawals from their accounts. At year-end 2000, rates are much lower when workers do not always find loans, or withdrawals has an effect on the income that is plan coverage, replacement rates from projected 401(k) account in equity funds relative to the average at all Loan Activity in 1998,” projected to replace at least 37 percent of their income shortened the vesting schedule for matching contribu- about 6- ICI Perspective, Vol. 6, no. 1; borrowed from their accounts, the projected proportion of Although there is no systematic variation in median accounting, actuarial, or other such professional advice. income. account balance, asset allocation, loan balance, and 65, over which time defined contribution plan assets in activity inside 401(k) plans and behaviors at job 42 always have contributions (their own and/or employer Accumulation Projection Model, while an Appendix, living. See Steuerle, Spiro, and Carasso (May 2000) for a This baseline assumption is based, in part, on Ippolito Quartile at Age 65 ....................................................... 65 5 a reserved. 50 a on recent information from a very large and representa- equity returns would be similar to historical returns accumulations into an IRA, typically at job change. Thus, 17 percent. returns, respectively (Figure 8). Effect of Participant In the bear market careers are projected to be pulled down between 2.9 and income benefit 1 for the first year of retirement was Social Security 401(k) Accumulation Social Security and 401(k) Accumulation officers, trustees, or other sponsors of EBRI, University of Michigan. However, the asset allocation of the IRA changes over contributions into their 401(k) accounts in that year. In McCarthy, Mike, and Liz McWhirter. “Are Employees Choices About Retirement.” Washington, DC: rollover IRAs or accounts left at previous employers. individuals might end up if they continue their current pared with the baseline scenario (Figure 7). tion, the historical total returns of the S&P 500 from the Health and Retirement Study. to replace 27 percent of preretirement income with their Institute and Employee Benefit Research Institute, sure to 401(k) plans could generate in retirement for an total return history. Alternatively, the model is then Sabelhaus, John, and David Weiner. “Disposition of projected effects of the timing of relatively brief but EBRI/ICI model projected a full career for them. Older 2001), Munnell, Sundén, and Taylor (December 2000), EBRI/ICI participant data indicate that only 4.5 percent ages, while still rebalancing over time away from equity with 401(k) accumulations, half are projected to replace income replaced at retirement would have been slightly retirement caused projected median replacement rates to participants’ behaviors over time in the model are replacement rates between soon-to-retire participants The term 401(k) accumulations covers 401(k)-related withdrawal information for each participant drawn from themselves in 401(k) plans (see Figures 1 and 7). would benefit from additional contributions as well as projected to be provided in retirement by 401(k) accumu- accumulations fall drastically. Among those reaching age and tions in defined contribution plans. EBRI Issue Brief no. 218 (Investment Company change over the remainder of their careers (Figure 2). discussion of replacement rates. Figure 2, Diagram of Basic Elements of the EBRI/ICI contributions) every year raises projected replacement which presents the details of the model, elaborates on EBRI Issue Brief is registered in the U.S. Patent and Trademark Office. ISSN: 0887-137X 0887-137X/90 $ .50+.50 (1997), which suggests that employers use defined tive sample of 401(k) participants. Furthermore, unlike experienced by the S&P 500 from 1926 to 2001. The S&P • Another projection scenario forecasts participants experiencing a simulated three- to paint the complete picture of 401(k) plan savings, both 2 36 scenario, participants’ equity holdings (equity funds, calculated for each participant at age 65. By design, Quartile 1 Quartile 2 3.7 percentage points (depending on income quartile) Quartile 3 Quartile 4 EBRI-ERF, or their staffs. Neither EBRI nor paths. The typical individual 401(k) participant behav- participants at year-end 2000 in the EBRI/ICI database time as the participant ages just as the 401(k) account is addition, among those with contributions, the total Ann Arbor, MI. Available: http:// May 17–18, 2001. Selecting a sample of participants of all ages and tenures be between 16.0 and 21.0 percentage points (depending beginning of 1926 to the end of 2001 were used. balances whether maintained as balances remaining in and U.S. General Accounting Office (October 1997). Missing the Big Picture? Study Shows Need for For of participants had taken a withdrawal during the year. 401(k) accumulations, while half of the highest income securities. Institute, January 2000, and Employee Benefit January 1999). at least 52 percent, and one-quarter are projected to individual, most of the discussion in this report focuses used to project the effect of concentrated negative (bear highly concentrated bear and bull markets are also Lump-Sum Pension Distributions: Evidence from Tax higher. For example, participants in the lowest income presented in the appendix (see endnote 12). and those not retiring for many years, there is variation the year-end 2000 EBRI/ICI database, 401(k) accumula- investment returns. In addition, it appears that for some lations. In addition, changing assumptions regarding the 65 between 2030 and 2039, the lowest income quartile at 12 50 65 Bibliography rates, but not by much compared with the importance 14 the discussion of the section. The model forecasts for a 401(k) Accumulation Projection Model ...................... 7 Technically, this is called the primary insurance In all other remaining years of the model, partici- contribution plans to sort workers, specifically to attract 7 500 total equity returns used in the analysis are from year bear market (negative equity returns) either early in their careers, near the household survey information, which can suffer from Quartile 1 Quartile 2 compared with the baseline scenario. A market down- rollover IRAs that were generated solely from 401(k) Quartile 3 Quartile 4 company stock, and the equity portion of balanced funds) Social Security replaces a higher percentage of lower- For simplicity, in this study, it is assumed that all EBRI-ERF lobbies or takes positions on specific Projected Replacement Rates From Even, William E., and David A. Macpherson. “Sex contribution averaged 9.3 percent of salary but varied Behavior on Projected hrsonline.isr.umich.edu/ Accessed 10/23/2002. iors observed today are derived from an analysis of each year in the model, each participant holding equity Ongoing Financial Education.” Benefits Quarterly. Venti, Steven F., and David A. Wise. “Choice, Chance, quartile are projected to replace 41 percent of salary or on the projection model experience of participants who market) and high (bull market) equity returns on Returns.” National Tax Journal. Vol. LII, no. 3 do not have as much time before retirement as younger rebalanced over time. In addition, the IRA assets are in the analysis understates the 401(k) accumulation on income quartile) higher compared with the baseline. the current and past employer plan(s) or as rollover IRA Younger participants, who may be penalized for with- influence of job changes also has a significant effect on age 65 are projected to replace 27.5 percentage points Research Institute, February 2000). replace 71 percent (Figure 6, bottom panel). Among those examined. quartile at age 65, turning 65 between 2030 and 2039, of replacement rates within birth cohorts (genera- tions at the age of 65 are estimated for each participant. households defined benefit plan participants are pro- 24 57 32 Participant Activity Inside the 401(k) Plan of being offered a plan to begin with. 15 very large sample of 401(k) plan participants their future Figure 3, Median Replacement Rates From Social individuals who value saving because they tend to be CPS data for 1999, 2000, and 2001 were used. The See U.S. Department of Labor, Pension and Welfare amount (PIA). The PIA was calculated for the individual pants are projected to experience the normal distribution “Funds” include mutual funds, bank collective trusts, difficulties with participant recall, Ibbotson (2002). The Ibbotson series used is “large middle of their careers, or at the end of their careers. Forecasts of the effects of bear the EBRI/ICI data balance rollovers and 401(k) balances themselves are are projected to experience three contiguous years of an turn is projected to have a larger impact the closer it income individuals’ incomes than it does of higher- 51 25th Percentile Median individuals retire at age 65. policy proposals. EBRI invites comment on this 75th Percentile U.S. Bureau of the Census. projected to earn the same rates of return as the 401(k) with age, tenure, and salary. For each year in the model, Differences in Retirement Income: Recent Trends and replacement rates at retirement for those participants 2.5 million 401(k) participants drawn from the year-end securities is randomly assigned a rate of return from the Vol. 16, no. 1 (First Quarter 2000): 25–31. Survey of Income and more (Figure 4, top panel). Among participants reaching and Wealth Dispersion at Retirement.” were in their late 20s in 2000 and who are projected to participants. In addition, it examines the impact of the (September 1999): 593–613. NBER Work- participants and therefore experience a shorter time in drawals, were less likely to have taken a withdrawal in this birth cohort, but in the highest income quartile at are projected to replace 0.4 percentage points more of balances (generated at job change). tions). This variation reflects the fact that individuals Hurd, Michael, Lee Lillard, and Constantijn Panis. Using current life expectancies and discount rates, these jected to have continuous defined benefit coverage at the results at retirement. Figure 7 presents the effects of less of preretirement income while those in the highest 401(k) Accumulations An Did you read this as a pass-along? Stay ahead of employee benefit Effect of Investment • The model simulations show that participant activi- Security for Participants Turning 65 in the Year behavior inside 401(k) plans as well as over their careers Benefits Administration (Winter 2001–2002), which life insurance separate accounts, and other pooled highly productive workers. CPS is a monthly survey of about 50,000 households Could we send a friend or colleague a complimentary participant’s earnings history and did not consider the of returns historically seen in the United States between used to construct the 401(k) plan behaviors in the model company stocks total returns.”markets on 401(k) balances show that a bear market in equities is projected to have included in “401(k) accumulations” in this analysis. 22 annual –9.3 percent total return, which reduced equity income individuals. Among participants who were in occurs to retirement, even though older participants tend research. 2000 database collected by the Employee Benefit Re- Source: Tabulations from EBRI/ICI 401(k) Accumulation Projection Model. the model. Because some of these older participants may account assets. it is determined whether each participant had contribu- Program Participation Future Prospects.” who are older and with lower tenure at year-end 2000. Mitchell, Olivia S., P. Brett Hammond, and Anna M. historical range of returns. Company stock (the plan Working Paper . Available:. Tallahassee, FL: compared with participants in their 60s. In the model, age 65 between 2020 and 2024, half of the lowest income Analysis of the Choice of Cash Out, Maintain, or ing Paper age 65, three-quarters are projected to replace at least reach age 65 between 2035 and 2039. For comparison, 8 timing of the bear or bull returns—early in an Samwick, Andrew A., and Jonathan Skinner. “How Will Once employed and eligible to participate in the 401(k) . No. 7521. Cambridge, MA: National EBRI Issue Briefs for only $49/ their preretirement income compared with the baseline are likely to differ in their experience in the work force accumulations are then expressed as an annual income— same employer for the remainder of their careers, which changing some key 401(k) participant behaviors and income quartile are projected to replace 39.4 percentage issues with your own subscription to Replacement Rates at Participant Behavior Over Working 2 The 401(k) accumulations are converted into an a ties such as taking loans, taking preretirement as they change jobs. Projected replacement rates are Indicated, by Income Quartile at Age 65 reports that 27 percent of private wage and salary For other examples of model simulations, see copy of EBRI Issue Brief? ................. 11 investments. conducted by the Bureau of the Census for the Bureau of The 401(k) accumulation includes 401(k) balances at employer(s) and rollover IRA balances. possibility of a spousal benefit, which can be substan- 1926 and 2001. There was no adjustment to the average the largest effect the closer it occurs to age 65 (retirement), even though older are based on administrative records. 43 3 to have diversified their portfolios away from equities. If Using current life expectancies for individuals age 65 holdings by 25 percent by the end of the three-year their late 20s at year-end 2000, half of those in the Ameriks, John, and Stephen P. Zeldes. “How Do House- Another source of income in retirement is private 53 12 Source: Tabulations from EBRI/ICI 401(k) Accumulation Projection Model. year electronically e-mailed to you or $99/year printed and mailed. 18 tions to his or her account based on the percentage of For example, see Copeland (July 2002); Fidelity www.sipp.census.gov/sipp/ Accessed 10/23/2002. Florida State University, Department of Economics, search Institute (EBRI) Thus, a second sample was carved out of the first—a Rappaport. Forecasting Retirement Needs and Retire- and the Investment Company quartile at age 65 are projected to replace at least 43 per- Bureau of Economic Research, February 2000. this study estimates Social Security replacement rates (a individual’s career, near the middle (age 39 to 41), or plan, the participant must decide whether to contribute Defined Contribution Pension Plans Affect Retirement have recently changed jobs and left account balances at a sponsor’s stock) was modeled to experience a wider range each year it is determined whether each participant experiences. points less, compared with the baseline model (Figure 7). Annuitize Pension Rights At Job Change or Retire- 54 percent of their preretirement income using their Projected median replacement rates from scenario (Figure 7). Similarly, for the highest income The appendix is available through ICI’s Web site at and/or in their 401(k) plans. an annuity or set of installment payments. would tend to overstate defined benefit plan benefits The income a income stream—an annuity or set of installment pay- withdrawals, or cashing out account balances at job reported for 401(k) participants retiring in the near and Career Figure 4, Median Replacement Rates From 401(k) VanDerhei and Copeland (July 2002); Poterba, Venti, Labor Statistics (see the bibliography for the Web site). workers were active participants in defined contribution Baseline model assumes participants have continuous coverage in 401(k) plans. tially larger than an individual’s own benefit in some historical experiences to offset or otherwise account for Returns on Projected 4 participants typically have diversified their portfolios away from equities. A three- and a discount rate, the accumulations at retirement are period. In the bull market scenario, participants’ equity Much of the previous simulation model research the bear market is projected to occur immediately before 33 For more information about subscriptions, visit our Web site at lowest income quartile at age 65 are projected to have pension savings. Although workers may be offered b hold Portfolio Shares Vary With Age?” Columbia U.S. Department of Labor. Bureau of Labor Statistics, IRA Withdrawals participants with that age, tenure, and salary level from Investments (2001); Hewitt Associates (May 2000); December 2001. “high-tenure sub-sample”—of nearly one million partici- Institute (ICI) of returns to capture the higher variation one stock ment Wealth—IRA owners may choose to take . Philadelphia, PA: University of in their collaborative effort known as the Wise, David A., ed. cent of salary, while half of the highest income quartile retirement income most Americans are likely to receive) immediately prior to retirement (age 63 to 65). to the plan, and his or her employer may make contribu- Generally, equities represent about 60 percent of Income?” Working Paper Facing the Age Wave . Hanover, NH: Dartmouth . Stanford, CA: previous employer or rolled them into an IRA, they may takes a withdrawal based on the percentage of partici- 401(k) accumulations, half are projected to replace at www.ici.org. Hard copies may be obtained from ICI’s quartile, it is projected that 0.3 percentage points more Send an issue to The 401(k) accumulation includes 401(k) balances at employer(s) and rollover IR stream provides a means of comparing income generated 401(k) accumulations for participants reaching age 65 particularly for those in final average plans. U.S. Depart- It is important to note that these reductions may be ment. RAND Institute Mimeo. Santa Monica, CA: The A balances. ments—using current life expectancies at age 65 and Retirement change reduce projected 401(k) accumulations and Accumulations for Participants Turning 65 in the distant futures. For comparison, replacement rates for plans only, 15 percent were active participants in both and Wise (November 2001); Samwick and Skinner The survey has been conducted for more than 50 years. cases. The PIA calculated for each individual is the sum the concentration of the three good or the three bad www.ebri.org or complete the form below and return it to EBRI. has been aimed at comparing defined benefit and defined year bear market for those early in their careers would reduce median replacement converted into an annual income stream—an annuity or 10 holdings are projected to experience three years of an Social Security replace 48 percent or more of their retirement, projected median replacement rates fall defined benefit and/or defined contribution pension University Working Paper. New York: Columbia EBRI/ICI Participant-Directed Retirement Plan Data balanced funds’ asset holdings (see ICI, Quarterly ________. have 401(k) accumulations not included in their EBRI/ withdrawals from their IRAs, although withdrawals year-end 2000 who had a contribution. Once it was McCarthy and McWhirter (2000); Burman, Coe, and and U.S. Bureau of the Census. pants, who have long tenure for their age group at experiences compared with a market average. All other Pennsylvania Press, 2000. The Impact of Rising 401(k) Pension Coverage Current Population pants with that age, tenure, and salary that had a are projected to replace 60 percent of salary or more RAND Institute, October 1998. Hoover Institution Press, Stanford University, 1997. least 69 percent, and one-quarter are projected to replace in addition to 401(k) accumulation replacement rates. tions. Over the year, a participant may decide to borrow College, October 2001. Research Department. of income would be replaced. These small effects do not Social Security Replacement Rates—Within any given in the first year of retirement to income estimated to between 2030 and 2039 are compared with the baseline ment of Labor (DOL), Bureau of Labor Statistics partially offset by participation in a defined benefit plan Job Change—Workers often change jobs over the course Organization discount rates. The replacement rate compares the thus replacement rates at age 65. Because loans are Social Security, a retirement income most workers are 47 Year Indicated, by Income Quartile at Age 65 defined benefit and defined contribution plans, and ........ 12 (October 2001); VanDerhei and Copeland (April 2001); The CPS is the primary source of information on the of three separate percentages of portions of their average years imposed on the participants. rates from 401(k) accumulations by an estimated 2.9–3.7 percentage points, com- contribution plan results. For example, VanDerhei and Conclusion 5 set of installment payments —for comparison with annual 31.2 percent total return, which increased equity preretirement income when they reach age 65 between Replacement Rates between 13.4 and 17.7 percentage points (depending on benefits, this report focuses on 401(k) plans. At age 65, This document is available electronically in pdf University, December 3, 2001. determined that a contribution occurred, the total Survey on Future Pension Income. Montalto, Catherine P. “Retirement Savings of American Collection Project. . Available: www.bls.census.gov/cps/ Report submitted to Wolff, Edward N. (middle panel). However, these relatively low replace- Supplemental Data Scholz, John Karl. “Can Americans Maintain Pre- While it is impossible to anticipate every possible Effect of Projecting Total Returns to Equities Based on Retirement Insecurity: The Income ). ICI year-end 2000 account balance. Missing these prior taken prior to age 59-1/2 may face a penalty tax in Gale (September 1999); Yakoboski (1999 and October year-end 2000. Results for this sub-sample of high- investments earned a projected nominal total return of Ibbotson Associates. model experience for those participants. withdrawal in 2000. If it is decided that the participant or a non-401(k)-type defined contribution plan, or with 89 percent or more. A similar distribution of replacement from his or her 401(k) account or take a withdrawal. SBBI (Stocks, Bonds, Bills, and account for the possible influence of the change in loan birth cohort, there is variation in projected Social have been earned during the participant’s working (September 19, 2002), reports that only a little over one- of their working careers and participants in the EBRI/ Name income or installment payment generated in the first 13 forecast to be paid back to the account in full, their Address likely to receive, are also projected. After examining the Figure 5, Distribution of Replacement Rates From Poterba, Venti, and Wise (August 1999); and Even and labor force characteristics of the U.S. population. 7 percent were in private defined benefit plans only. indexed monthly earnings (AIME). The portions depend Several EBRI and ICI members provided records on pared with 13.4–17.7 percentage points for those immediately before retirement. income prior to retirement. Copeland (April 2001) forecast that an increasing holdings by 126 percent. Participants continue to be 37 income quartile) compared with the baseline. 2035 and 2039, while half of those in the highest income the model forecasts that 401(k) accumulations—the sum Ameriks, John. “Trends in TIAA-CREF Participant format. To order online, visit www.ebri.org/ 19 addition to ordinary income taxes. Very few young contribution rate is calculated for the participant using a 1999); Sabelhaus and Weiner (September 1999); Poterba, cpsmain.htm Accessed 10/23/2002. Department of Labor, Pension and Welfare Benefits tenure participants are also analyzed. 5.3 percent. Households: Asset Levels and Adequacy.” Analysts often evaluate the adequacy of an Report to ment rates are in part the result of not correctly Shortfalls Awaiting the Soon-to-Retire scenario, several variations in both participant behaviors 1929 to 1978 Historical Returns Finally, asset allocation (which changes with age) and Retirement Consumption Standards in Retirement?” —Projecting equity total . Washington, accumulations understates the replacement rates takes a withdrawal, a regression equation is used to rates occurs among participants projected to retire in usage on other activities, such as contributions (which Security replacement rates and the variation is wider in career immediately prior to turning 65. quarter of older workers (age 55 and older) have long contributions to traditional IRAs during the periods of Inflation) 2002 Yearbook: Market Results for 1926- ICI model do so as well. Based on SCF job duration The ratio of 34 The results discussed above represent the typical Organization year of retirement to the final five-year average effect on replacement rates at retirement in the model Social Security for Participants Turning 65 in the projected replacement rates across individuals retiring at Generally, fixed-income securities represent about active participants in 401(k) plans they administered at the effect of each participant behavior or activity on Macpherson (March 1998). City/State/ZIP on the year in which the worker reaches retirement. For returns on replacement rates at retirement. The final 54 proportion of retiree wealth (and therefore retirement Combined, income from Social Security and 25 58 exposed to a random draw from historical equity market quartile are projected to experience a Social Security 56 always of balances in 401(k) plan accounts and rollover IRA have contributions to their plans every year pants do not make contributions themselves or have publications or call EBRI Publications at (202) Premium and Asset Allocations: 1986–2000.” Research individual’s retirement savings by comparing the income The EBRI/ICI 401(k) model does not allow for con- U.S. Department of Labor. Bureau of Labor Statistics. potentially to be experienced by these older participants. individuals withdraw from their IRAs; however, among regression equation. Contributions are limited by Venti, and Wise (August 1999); Hurd, Lillard, and Panis Administration. March 1998. For a complete discussion of the IRC regulations the Consumer Federation of America and In any given year, equity and company stock A brief description of the key elements of the estimate the percentage of the account balance with- accounting for accumulations that occurred at employers 2001 DC: Economic Policy Institute, 2002. other years (Figure 6, top and middle panels, for and equity market returns are analyzed. returns based on the worst 50-year period tracked by the investment returns must be accounted for. The projec- Working Paper . Chicago: Ibbotson Associates, 2002. . Madison, WI: University of Wisconsin, might be reduced) lower income quartiles. For example, among participants and withdrawals (which might be retirement income to preretirement income—known as a Effect of Contributions— tenures (20 years or more) at their current employers lapses in 401(k) coverage. However, IRA contribution behavior, EBRI/ICI model participants may change jobs; Based on contribution activity • Similarly, a simulated three-year bull market (positive equity returns) is projected to In defined benefit plans, participants primarily bear experience of 401(k) participants projected to work a preretirement income. The 401(k) distributions are not is the smallest. different future dates, variation within a group retiring 40 percent of balanced funds’ asset holdings (see ICI, Year Indicated, by Income Quartile at Age 65 ........ 14 Address year-end 2000. These plan administrators include projected replacement rates at retirement, while the example, for 2002 the PIA was 90 percent of the first section contains a brief conclusion. income) is expected to be managed by retirees as a result 3 Your Name 401(k) plan accumulations at retirement is projected to 29 returns (based on S&P 500 total returns from 1926 replacement rate of at least 15 percent (Figures 1 and 3). Current retirees have not had access to 401(k) plans long Effect of Three-Year Bull Market— raises the percentage of income that is projected to be balances—are projected to generate income to replace a 659-0670. Dialogue. No. 65. New York: TIAA-CREF Institute, Figure 9 also presents Federal Reserve Board. Internal Revenue Code (IRC) regulations and those “Employee Tenure Summary.” News Release. that retirement savings are projected to generate in holdings each earn a randomly selected rate of return governing 401(k) plan participant contributions and a employer contributions to their 401(k) plans in any given tributory traditional IRAs or for participation in defined The Employee Benefit Research Institute is a non- DirectAdvice.com. Columbus, OH: The Ohio State Survey of Consumer Finances. Yakoboski, Paul. “Rollover Rates Continue to Rise.” previous to the EBRI/ICI participant’s year-end 2000 S&P 500 reduces projected median replacement rates Department of Economics and the Institute for If equity and bond markets provide returns that Thus, as explained earlier, a high-tenure sub-sample of the few younger individuals taking withdrawals, the (October 1998); Yakoboski (August 1997); Poterba, Venti, participants’ behavior in the model is presented in this Internal Revenue Service. observed among EBRI/ICI 401(k) participants, the drawn. limits are lower than those permitted in 401(k) plans example). tion model’s treatment of these activities inside the Publication 575 Pension and increased). reaching age 65 between 2035 and 2039, three-quarters replacement rate (data for January 2002), suggesting the reward of long however, they were projected to have continuous careers —serves as a rough indicator of whether have a larger positive effect on projected account balances and replacement rates the “employment” and “employer” risk. Employment risk indexed for inflation over retirement, whereas Social continuous career and always at employers offering a at the same time is discussed. The report then analyzes Quarterly Supplemental Data). Figure 6, Distribution of Replacement Rates From $592 of their AIME plus 32 percent of their AIME over mutual fund companies, insurance companies, and following section focuses on the effect of investment replace a substantial proportion of individuals’ City/State/ZIP of participation in defined contribution and individual through 2001) for every other year in the model, while 23 46 the projected effect of the timing of a bull market in The EBRI/ICI model preserves Social Security substantial portion of projected preretirement income. 58 year. Altering this assumption so that participants enough to examine the ability of such plans to generate replaced at retirement. Because lower income partici- October 2000. 41 amount withdrawn is sizable. Available: www.federalreserve.gov/pubs/oss/oss2/ 19 September 2002. drawn from their respective distributions of the range of retirement to preretirement income. This “replacement section. profit, nonpartisan, public policy research organization detailed analysis of 401(k) plan participants’ contribu- University, April 26, 2000. Each 401(k) participant is projected to engage employer. When the high-tenure sample is analyzed, EBRI Mimeo. Washington, DC: Employee Benefit are near their historical norms, income generated from from 401(k) accumulations by 10 to 13 percentage points 401(k) plan is summarized below. Research on Poverty, August 13, 2001. participants is also analyzed. of those in the lowest income quartile at age 65 are retirees are expected to be able to maintain their baseline model assumes that a small portion of partici- tenure in traditional defined benefit plans does not and have no possibility of employer contributions. Annuity Income (i.e., they were always employed). . Washington, DC: U.S. Department of Under baseline closer it occurs to retirement. 401(k) plan (the baseline model). However, changing occurs because a participant might not stay in the job Security benefits are. In addition, if the participant Mail to: EBRI, 2121 K Street, NW, Suite 600, Washington, DC 20037 Mail to: EBRI, 2121 K Street, NW, Suite 600, Washington, DC 20037 or Fax to: (202) 775-6312 or Fax to: (202) 775-6312 EBRI Issue Brief Number 251 • November 2002 • © 2002. EBRI 2 20 28 10 8 4 12 18 26 16 14 22 24 6 November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief November 2002 • EBRI Issue Brief 25 19 15 11 23 21 13 27 17 1 5 7 9 3 Issue Brief Issue Brief Participant Behavior Over Participant Activity Inside the 401(k) Plan Working Career

Volume 251

Pages 28

EBRI Issue Brief

Sarah Holden

Jack VanDerhei

Retirement