- A rapidly growing public policy concern facing the United States is whether future generations of retired Americans, particularly those in the “baby boom” generation, will have adequate retirement incomes. One reason is that Social Security's projected long-term financial shortfall could result in a reduction in the current-law benefit promises made to future generations of retirees. Another reason is that many baby boomers will be retiring with employment-based defined contribution (DC) plans, as opposed to the “traditional” defined benefit (DB) plans that historically have been the predominant source of employer-provided retirement income.
- These factors are likely to reduce the amount of life annuity benefits that future retirees will receive relative to current retirees, raising questions as to whether other sources of retirement income--such as individual account plans (DC plans and individual retirement accounts, or IRAs)—will make up the difference.
- This Issue Brief highlights the changes in private pension plan participation for DB and DC plans and provides some possible explanations for these changes. Results are presented from the Employee Benefit Research Institute's (EBRI) Retirement Income Projection Model that quantify how much the importance of individual account plans is expected to increase because of these changes. This Issue Brief also discusses the risk of outliving one's assets, since a greater fraction of pension wealth is projected to come from “nonguaranteed” sources.
- Results of the model are compared by gender for cohorts born between 1936 and 1964 in order to estimate the percentage of retirees' retirement wealth that will be derived from DB plans versus DC plans and IRAs over the next three decades. Under the model's baseline assumptions, both males and females are found to have an appreciable drop in the percentage of private retirement income that is attributable to defined benefit plans (other than cash balance plans). In addition, results show a clear increase in the income retirees will receive that will have to be managed by the retiree. This makes the risk of longevity more central to retirees' expenditure decisions.
- The implications of these model results for retirees are significant. First, individuals—rather than the pension plan sponsor—increasingly will have to manage their retirement assets and bear the risk of investment losses. Second, since most retirees' non-Social Security retirement income will be distributed as a lump sum or in periodic payements (from a defined contribution plan or IRA) rather than as a regular paycheck for life (from a defined benefit plan), retirees will need either to purchase an annuity from an insurance company or carefully manage their individual rate of spending in order to avoid outliving their assets.
EBRI Issue Brief
PDFTextVersion is null
Jan. April 2001 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 rate paid in Washington, DC. POSTMASTER: Send address changes to: EBRI Issue Brief, 2121 K Street, NW, Suite 600, Mar. Chart 4 Table 3 Chart 2 EMPLOYEE Chart 1 Chart 9 Washington, DC 20037-1896. Copyright 2001 by Employee Benefit Research Institute. All rights reserved, No. 232. Chart 7 Jack VanDerhei, Temple University and EBRI Fellow, and Craig Copeland, EBRI, wrote Average 401(k) Account Balance by Age and Tenure, 1999 Percentage Change in Composition of Estimated Future r Chart 3etirement The Changing Face of Private The main result of the Percentage of Households by Pension Type, Composition of Estima Estimated Percent ted Retirement Weal age of Private W th for Females a age and Salary W t Social Security Normal Retirement A orker Participation ge BENEFIT Predicted 401(k) Contributions for Selected Persons and Plans this Issue Brief with assistance from the Institute’s research and editorial staffs. Any views Private Trusteed Pension Plan Assets Vs. Individual Retirement Account Assets, 1992-1999 Apr. Retirement Wealth at Social Security Normal 1992–1998 Rates Under Primary and Supplement Under Baseline Assumptions, al Pension Pl by Birth Cohort ans, 1977–1996 $250,000 RESEARCH 12% income will depend less expressed in this report are those of the authors and should not be ascribed to the officers, Retirement Age for Various Scenarios model developed in this The Employee Benefit Research Institute (EBRI) was founded in 1978. Its mission is $2.50 60% $2.45 100% 40% Age 52.7% 57.3% $2.47 INSTITUTE ho we areRelative to the Baseline Assumptions: May ® to contribute to, to encourage, and to enhance the development of sound employee benefit trustees, or other sponsors of EBRI, EBRI-ERF, or their staffs. Neither EBRI nor EBRI- W 1992 Defined Benefit Defined Contribution Individual Retirement Accounts $2.21 20s on “guaranteed” forms of 1964 Birth Cohort, by Gender Issue Brief is that the 50% Retirement Plans $200,000 programs and sound public policy through objective research and education. EBRI is the only private, nonprofit, 10% ERF lobbies or takes positions on specific policy proposals. EBRI invites comment on this 35% Primary Defined Benefit Plan 1995 $2.03 90% $2.03 30s 39.6% $1.99 37.6% Jun. nonpartisan, Washington, DC-based organization committed exclusively to public policy research and education on $1.91 $2.00 40% 1998 benefits (such as DB plan research. 40s expected increase in the 30% $1.78 economic security and employee benefit issues. EBRI’s membership includes a cross-section of pension funds, Type Male Female 80% $1.73 8% 50s 25.6% by Jack V $150,000 anDerhei, Temple University and EBRI Fellow, 30% 22.8% 21.8% 22.8% Jul. businesses, trade associations, labor unions, health care providers and insurers, government organizations, and $1.59 annuities and Social $1.55 19.9% 25% Supplemental Defined Contribution Plan(s) need for income com- 60s $1.47 Panel B: Cash Balance Plans Double service firms. 20% 70% $1.50 $1.40 $1.32 Defined benefit –4.1% –2.6% $1.29 6% $1.25 Security) and more on 20% Aug. and Craig Copeland, EBRI $100,000 $1.21 ing from sources that Defined contribution + cash balance 10% 2.9 2.9 $1.15 $1.07 $1.09 60% 100%, up to 5% $1.06 Individual retirement account 0.3 0.0 15% EBRI’s work advances knowledge and understanding of employee benefits and their $0.99 $0.95 DC plan benefits—which 0 are not guaranteed for Sep. $1.00 75% up to 6% hat we do 4% $0.87 W 39 importance to the nation’s economy among policymakers, the news media and the public. 43 Defined Benefit Defined Both will be determined, in large part, by individual decisions for females. IRAs were assets. Furthermore, since annuities would provide a tion to younger workers are usually higher under DC and Challenges for Workplace Savings in America increasing attention paid to these issues since passage of (1998) 50% $50,000 Employee Benefits in Medium and Large. by the employee (although the employee may have the picture of the implications for the baby boom generation. administrative costs had grown to approximately these changes. Because individual account plans tend to 2000): 112–124. sions.” National Tax Journal (and between approximately 27 per- •Similar figures for females are ASince this was the most current three-year final-average plan an estimated increase of (June 1998): 263–289. three years of tenure with their current employer to in the year 2000 is expected to live another 16.4 years (to SCF data holds that likely will rely primarily on DC plans for decision to annuitize one’s assets is A combination of Form 5500 to estimate the relative likelihood that the A rapidly sions is rather small. appear from this limited sample that well over 50 for the 1964 birth cohort: 41.6 percent for males and does break out source of contributions (e.g., The American Council of Life Panel C: Assuming a Less Aggressive Tables and Charts 10% 50% up to 6% Only Contribtuion Only It does this by conducting and publishing policy research, analysis, and special reports on employee benefits issues; will be determined, in 4 22 life. Consequently, provided in panel A of table 2. plans than they are under traditional DB plans. This is guaranteed payment, individuals would be protected (e.g., whether to contribute to the plan, how to invest, estimated to decrease by Boston, MA: Fidelity Investments, 2001. the SAVER Act in 1997, much remains to be done. This Private Establishments, 1995. Long Term Asset Allocation option of converting the distributions to an annuity). Mitchell, Olivia S., James M. Poterba, Mark Our model estimates that for today’s retirees with either 210 percent of the cost of administering a 401(k) plan. Brown, Jeffrey R. “Are the Elderly Really Over- pay lump-sum benefits cent and 55 percent of females age with a PIA offset. number available, it was used in at retirement rather than life 5.2 percent, while their age 81), while a female is expected to live to about age $198,595 for participants in their 60s who have been data and self-reported results was also virtually irrevocable; otherwise the retirement income. But, these higher DC participation balances are rolled over to an IRA, left with the previous growing public percent of the employee contributions occur at “corner Insurers (2000) reported that 2.75 million individuals 40.0 percent for females. The differences between the employee before-tax, employee after-tax, Oct. • A rapidly growing public policy concern facing the United States is whether future holding educational briefings for EBRI members, congressional and federal agency staff, and the news media; and 75% up to 2%, 50% for 3–5% Plan Type 40% Primary Defined Contribution Plan Table of 2% 5% Defined benefit 5.9 4.9 Chart 1, Estimated Percentage of Private Wage The same information is pre- Gale, William, Leslie Papke, and Jack VanDerhei. $0.50 with the current employer for at least 30 years (thereby from imperfect decisions on mortality when making 85, an additional 19.6 years of life. Yet, approximately 3.3 percent and 3.7 per- how to manage plan benefits, and whether to preserve one reason that DC plans are seen as being more “por- used to estimate defined benefit rates are also accompanied by a dramatic increase in the report suggests that, at a minimum, financial issues in insurer would face a serious adverse- employer, transferred to a new employer, or used for www.bls.gov/special.requests/ocwc/oclt/ebs/ policy concern points” that would not be identified in the data described were covered by an individual immediate annuity policy defined benefit, defined contribution and/or IRAs, Hustead found that these relative administrative cost model and the set of assumptions utilized in Poterba, annuities, this Warshawsky, and Jeffrey R. Brown. “New Evidence Annuitized? New Evidence on Life Insurance and The notion of cash balance plans presents at employer matching, qualified non-elective Issue Brief 65), depending upon the equity • Athe baseline estimates. However, cash balance plan; or concludes by discussing the female counterparts are 0 large part, by individual sponsoring public opinion surveys on employee benefit issues. EBRI’s Education and Research Fund (EBRI-ERF) generations of retired Americans, particularly those in the “baby boom” generation, Source: Employee Benefit Research Institute estimates of the 1992, 1995, retirees will be at a Defined contribution + cash balance –0.8 –2.1 0% 31 and Salary Worker Participation Rates Under Nov. 0–2 Years >2–5 Years >5–10 Years >10–20 Years >20–30 Years >30 Years and 1998 Survey of Consumer Finances. “Understanding the Shift Toward Defined Contribu- 0 least a conceptual difficulty for purposes of this distinc- on the Money’s Worth of Individual Annuities.” allocation, would run out of resources sented graphically in charts 8 effectively eliminating any capability for IRA rollovers). table” from job to job. Another reason is that years of plan benefits upon job termination). However, although 12 percent of the males and 8 percent of the females who cent for males and expenditure choices. This would also potentially limit the participation models; however, it selection situation, making it nearly retirement must become a higher education priority in percentage of households that may have DC plans as other purposes. These probabilities are used to estimate ebb10015.pdf.performs the charitable, educational, and scientific functions of the Institute. EBRI-ERF is a tax-exempt organization facing the above. in 1999. However, this is not too surprising considering approximately 39.0 percent of pension wealth for males increases disproportionately affected small plans. Venti, and Wise and those used in this report are far too risk of outliving one’s assets. Bequests.” contributions (QNECs), etc.), and a series of NBER Working Paper. • Aa significant amount of conver- flat-benefit plan. estimated to have a No. 7193. Cam- 30% Individual retirement account –3.3 –3.7 Introduction will have adequate retirement incomes. One reason is that Social Security’s pro- EB 0 RI 1977 1979 1981 1983 1985decisions. 1987 1989 1991 1993 1995 Contents Years of Tenure supported by contributions and grants. greater risk of outliving The American Savings Education Council Primary and Supplemental Pension Plans, (ASEC) and the Consumer Health the likely directions of these trends are known, little to females, respectively. costs to the public sector in the form of lower public service under DB plans with age and service require- ________. (1999). tion Plans.” Paper presented at the Brookings/ the future. Most Americans have relied on Social Secu- Employee Benefits in Medium and tion, since they are legally defined benefit plans but are would be available from defined benefit plans and American Economic Review bridge, MA: National Bureau of Economic Research, The two primary reasons that relative DB plan before death. (males) and 9 (females). It is sion activity has taken place since While the BLS information (December 1999): 1299– 4.9 percent increase. reach 65 will die before they reach 70, while about appears information in the latter is their impossible for the insurer to offer the baseline scenario in panel A of tables 1 and 2. Unfortunately, this database does not currently only source of employment-based retirement United States is the relatively high percentages of annuity payments numerous to mention; however, it is important to keep in In VanDerhei and Copeland (2001), preliminary computer algorithms has been developed to Age = 55 jected long-term financial shortfall could result in a reduction in the current-law Age = 40 Age = 30 Age = 35 Age = 25 Age = 22 $0.00 Dec. 20% Panel D: Assuming Terminated Vested Defined Education Council (CHEC) are programs of EBRI-ERF. They are coalitions of private- and public-sector institutions 40 Wage = $45,000 Wage = $25,000 Wage = $55,000 W 1977–1996 age = $20,000 .................................................................... Wage = $20,000 readily apparent from these Wage = $15,000 4 TIAA-CREF Institute/SIEPR Conference, ERISA after provide detailed information on other types of defined 17.5 percent of the males and 31.4 percent of the females Defined benefit plan date has been done to quantify the magnitude of these assistance program expenditures, as many individuals ments are not usually transferable from employer to rather unreliable with respect to annuities. Consequently, annuity rity to provide a basic monthly annuity for life, and the income. In a recent EBRI study of households’ pension Large Private Establishments, 1997. Source: T One way to demonstrate the first-order impact abulations from EBRI/ICI Participant-Directed Retirement Plan Data Collection Project. whether future often perceived by employees to be defined contribution findings are provided that introduce new methodology to retirees are presently receiving from private pensions (as 49.7 percent for females; defined contribution and cash administrative costs rose from 1981–1996 are the mind that their analysis was limited to the 401(k) 1318. 1999. classify additional plans by the types of provides significant detail on the that time. Elliot and Moore (2000) Lastly, the lack of understand- Defined contribution Although both EMPLOYEE 1992 1993 1994 1995 1996 1997 1998 1999 Source: Department of Labor, Pension and Welfare Benefits Administration, benefit promises made to future generations of retirees. Another reason is that many Private Pension Plan Bulletin: Abstract of 1996 Benefit Participants are Not Automatically Cashed Out their resources. Tenure = 25 Years Tenure = 10 Years Tenure = 5 Years Tenure =3 Years Tenure = 5 Years Tenure = 1 Year with the goals of public education on saving, retirement planning, health insurance, and health quality. 44 Chart 2, Percentage of Households by Pension Type, 25 Years: A Framework for Evaluating Pension Form 5500 Annual Report, no. 9 (Winter 1999–2000): 66. Mitchell, Olivia S., and James F. Moore. “Can Americans graphs that both genders have an contribution plans, nor does it allow analysis of defined benefits do not change who reach 65 will live until they reach 90. Other impor- employer. trends in terms of future retirement income. would have a guaranteed level of income for life above estimating current and/or future accrued benefits. life of a spouse. Should the nation move to individual participation rates from the Survey of Consumer Fi- of preretirement cashouts on eventual retirement income purchasers are faced with liquidity generations of retired Americans, particularly those in www.bls.gov/special.requests/ocwc/oclt/ebs/ plans in the way they accrue benefits. Since many of the expand the usefulness of modeling these data, as well as shown in charts 8 and 9) and by Social Security. balance plans would provide 33.2 percent for men and funding regulations and Pension Benefit Guaranty Year 1970s, and those requirements have also changed many ________. “How Should We Insure Longevity Risk in generosity parameters for defined benefit plans, prelimi- market and that they did not include the impact of the incentives provided to employees at various ing about annuities’ benefits appears report that 16 percent of the plan and IRA benefits do qualified DC Yet, Defined benefit 5.2 4.7 10% baby boomers will be retiring with employment-based defined contribution (DC) 2001 BENEFIT DB Vs. DC Introduction ..................................................................... 3 10 1 Defined contribution + cash balance 1992–1998 –1.8 .................................................................... –2.7 5 the eligibility levels for these programs that they may under this new assump- Reform, Washington, DC, September 17, 1999. accounts in Social Security in the future, these same ebb10017.pdf A model for quantifying the relative levels of cash balance benefits appear to be taken as lump sums, 32.5 percent for women; while IRAs would provide Corporation (PBGC) insurance premium requirements nary analysis indicated that several of these provisions Afford to Retire? New Evidence on Retirement Saving Pensions and Social Security?” to be a major reason for them not appreciable drop in the percent- pension plans among Fortune 100 not change under this An Issue in Brief, and DB plans no. 4 contribution balances that may have been left with tant considerations are medical advances from research Therefore, a database of defined benefit plan provisions nances (SCF), constraints in paying for large expendi- is to simulate the impact of a proposal that would force the “baby boom” generation, the percentage of households with only will have adequate retire- there are other reasons why retirees do not purchase life a better understanding of contribution behavior by times over their evolution. In contrast, DC plan regula- ancillary restriction on preretirement withdrawals on contribution levels (e.g., a 50 percent match plans, as opposed to the “traditional” defined benefit (DB) plans that historically Source: Jack L. VanDerhei and Craig Copeland, “A Behavioral Model for Predicting Employee Contributions to 401(k) Plans: Preliminary Results,” Source: Employee Benefit Research Institute, Pension Investment Report, 2nd Quarter 2000 (Washington, DC: Employee Benefit Research Institute, Chart 5 Individual retirement account –1.9 –2.9 DB Vs. DC Plans .............................................................. 3 EBRI Issue Briefs are monthly periodicals providing expert evaluations of RESEARCH 0% Chart 3, Private Trusteed Pension Plan Assets age of private retirement income Gustman, Alan L., and Thomas L. Steinmeier. “The such as the Human Genome Project and the increase in tion; however, the previous employers. The model used in this report uses retirement income sources is outlined below. In addition, not have had if they had not annuitized. VanDerhei, Jack L., and Craig Copeland. “A Behavioral for salary-related plans was constructed to estimate a DB plan decreased from 39.6 percent in 1992 to tures, particularly for those health-care related expenses all defined contribution plan funds not either retained in issues of asset management, rate of spending, and ment incomes. One reason is that Social Security’s North American Actuarial Journal, Vol. 5, no. 1 (January 2001): 80–94. the projected “balances” from these plans are added to annuities, which are central to any debate on the neces- 401(k) plan participants. A sequential response regres- 27.8 percent for men and 17.8 percent for women. For that Congress enacted during that period. Most DC plans tions are not as numerous, nor have they changed as being purchased. A study by the American Council on were likely to be highly correlated (especially for inte- IRAs. Adequacy.” (Center for Retirement Research at Boston College, for the first 6 percent of compensation). This The Journal of Risk and Insurance. companies in 1998 were cash new assumption; how- are tax-favored Vol. December 2000); and Craig Copeland, “IRA Assets Continue To Grow,” EBRI Notes, no. 1 (Employee Benefit Research Institute, January 2001): 1–8. Business Envir Comp onment any A Contribution Ra ur publications tes, by Age have been the predominant source of employer-provided retirement income. O 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 employee benefit issues and trends, as well as critical analyses of employee Participation Trends........................................................ 4 6 7 Vs. Individual Retirement Account Assets, Plans Stampede Toward Defined Contribution Plans: Fact or defined contribution plans for purposes of presenting the Life Insurers determined that consumer knowledge of 65, no. 3 (1998): 371–400. that is attributable to defined vehicles for relative importance of health care needs that are associated with old age. self-reported responses for whether an individual has a the results of the model are compared by gender for benefit accruals. 19.9 percent in 1998 (chart 2). Meanwhile, the percent- the previous employer’s plan or transferred to the new not covered by Medicare. Also, ordinarily the annuity financial literacy may apply to that program as well. projected long-term financial shortfall could result in a Model for Predicting Employee Contributions to On the other hand, annuitization also has Panel E: Assuming All Defined Contribution Plan Account benefit and having an inadequate retirement income. sity for, or mandating of, retirees to purchase life sion model was used to allow for the differing incentives the youngest baby boom males (born in 1964) this report are fully funded (by definition), while DB plans must frequently over their development. Plan sponsors point grated plans). Therefore, a time series of several August 2000). information has been used to expand the balance plans, and, that more ever, the relative plan contributions, any plan forfeitures, and invest- INSTITUTE 70% ® benefit policies and proposals. Each issue, ranging in length from 16–28 pages, thoroughly explores one topic. EBRI Why the Shift to DC Plans? ............................................ 5 63% Balances That are Not Retained in a Defined Contribution DC plans are thought to better match the new philoso- Max Match (5%) 1992–1999.................................................................... 7 defined benefit plans for cohorts born between 1936 and 1964 in order to estimate potential concerns. First, some individuals may be “over- Fiction?” This report makes it clear that the individual responsi- 401(k) Plans: Preliminary Results.” Industrial Relations. Vol. 31, no. 2 (1992): North American results in this report. Olsen, Kelly A., and Jack L. VanDerhei. “Defined estimates that 26.4 percent of their pension wealth will meet complex ERISA regulations to determine their Brown, Jeffrey R., and Mark J. Warshawsky. “Longevity- annuities was low: Consumers always focused how they hundred defined benefit plans per year was coded to benefit plans (other than cash generally, cash balance plans importance of defined providing defined contribution balance to estimate a participation Consequently, retirees will need to understand how age of households with only a DC plan increased from payments end when the annuity purchaser dies; thus, no employer’s plan at job termination to be rolled over to an reduction in the current-law benefit promises made to Self-reported results were used to initialize IRADefined Contribution + Cash Balance annuities from their defined contribution plan and IRA faced by the employees at various levels of contributions. to the growing administrative burdens and costs of previous sample and provides the predicted Defined Benefit The results ment returns on account assets. That is, employers do Individual Retirement Account Alternatively, DB plans tend to offer life annuities (a set Notes is a monthly periodical providing current information on a variety of employee benefit topics. • These factors are likely to reduce the amount of life annuity benefits that future EBRI’s Washington 58% 60% Plan on Job Termination Must be Rolled Over to an Work Force .................................................................. 5 Plan Max (9%) 2 28 phies of relationships among compensation, employee Chart 4, Average 401(k) Account Balance by Age balance). Females start with a 361–369. the 1964 birth cohort the percentage of retirees’ retirement wealth that will be annuitized,” meaning that they would be worse off from model, and the reported value is modeled as a function of much (or little) they can spend or to explore other accounts. Future IRA contributions were modeled from 37.6 percent to 57.3 percent, and the percentage of IRA on a mandatory basis. However, this provision alone bequest can be given. While one can annuitize only a bility model for retirement income will be tested in the future generations of retirees. Actuarial Journal The significant shift in participation rates from . Vol. 5, no. 1 (January 2001): 80–94 Another reason is that amount paid out regularly over time, typically monthly, account balances at retirement. Based on findings from 137 distinct matching formulae, be provided through defined benefit plans (a decline of funding status. In addition, virtually all private employ- operating a DB plan as a major disincentive to operating might die early (and lose their investment in the annu- allow for assignment to the individuals in the model. Contribution Plan Dominance Grows Across Sectors retirement income, they differ in a variety of important Insured Retirement Distributions from Pension Plans: employee contributions for 401(k) plans in have increased from 5 percent to benefit plans for the 1964 above show a not guarantee a specific benefit level to DC plan partici- Bulletin provides sponsors with short, timely updates on major federal developments in employee benefits. them to another qualified plan, and 37 percent rolled EBRI’s retirees will receive relative to current retirees, raising questions as to whether Individual Retirement Account (IRA) and Preretirement 402(g) Business Environment................................................ 6 performance, and profits (Campbell, 1996). For instance, and Tenure, 1999 19 ...................................................... 11 50% Fundamentals of Employee Benefit Programs offers a straightforward, basic explanation of employee benefit programs Estimating Current and Future Accrued Holden, Sarah, and Jack VanDerhei. “401(k) Plan Asset ity), instead of how they might live longer than expected and Employer Sizes, While Mega Defined Benefit ways. One is how plan contributions are made. Under a Although the Tax Reform Act of 1986 at least slightly higher defined benefit increases by 5.9 percent age and tenure similar to chart 4. avenues (such as the purchase of annuities) to reduce the derived from DB plans versus DC plans and IRAs over the annuitization. Retirees in poor health would be a VanDerhei, Jack L., Sarah Holden, and Carol Quick. Survey of Income and Program Participation (SIPP) households with both DB and DC plans remained steady would not be sufficient to ensure that the funds re- decades ahead. It appears that a tag line from the portion of one’s assets and leave the rest for bequests, DB to DC plans and the rise in the number of households many baby boomers will be retiring with employment- Withdrawals from an IRA are Prohibited for as long as the beneficiary lives), which beneficiaries a behavioral model was estimated that is able to control 32.4 percent), while their female counterparts will see ers that sponsor DB plans are required to pay premiums this type of retirement plan. Market and Regulatory Issues.” Mimeo. Boston, MA: The reasons for not purchasing annuities range this paper. 12 percent of all defined benefit birth cohort decreases by clear increase in pants, and benefits are directly related to investment 44% them to an IRA. Participants who rolled over their DC other sources of retirement income—such as individual account plans (DC plans and Source: Employee Benefit Research Institute, Retirement Income Projection Model. 41% 40% Regulatory Environment ............................................ 6 ®in the private and public sectors. Defined benefit The EBRI Databook on Employee Benefits –23.4 is a statistical reference volume on employee –20.5 increased competition, reorganizations, restructurings, Chart 5, Company A Contribution Rates, by Age concentration than men (49.7 per- ........ 11 for males and 4.9 percent the next three decades. Lastly, the sensitivity of these particular class of individuals that likely would be hurt Allocation, Account Balances, and Loan Activity in Choose to Save “401(k) Plan Asset Allocation, Account Balances, and education program will become more their defined benefit pension wealth fall to 37.2 percent, to the PBCG, which provides insurance upon plan (and gain the post-retirement lifetime income that the partially modified the constraints on integrated pension Plans Remain Strong: Where We Are and Where We DC plan, employer contributions are based on a prede- Harvard University, 2000. plans in just the previous two 1.8 percent and 1.9 per- risk of longevity. data, while future rollover activity was assumed to flow at 22.8 percent. These findings applied universally mained in the retirement system if individuals retained this still may not be as attractive to many individuals as based defined contribution (DC) plans, as opposed to the from rational financial choices to the lack of understand- for the tendency of employers to substitute between the Discussion Contribution behavior for defined the income returns. In comparison, employers offering DB plans with DC plans exclusively have also been accompanied are not responsible for managing. (However, lump-sum plan assets typically rolled over larger amounts: The 40% Benefits and Account Balances individual retirement accounts, or IRAs)—will make up the difference. Defined contribution + cash balance –23.4 –20.5 benefit programs and work force related issues. Changes in Asset Holdings ............................................. 6 45 5 and mergers have made businesses become leaner in Chart 6, Company A Contribution Rates by Salary .... 9 11 cent vs. 39.0 percent for the 1936 1999.” EBRI Issue Brief 28% no. 230 (Employee Benefit for females. by forced annuitization; in this case, their bad health Contribution Behavior estimates to key modeling assumptions is analyzed. from future separation from employment in those cases relevant with each passing year: Save now or work across household demographics. having access to all of their assets. the rights to withdraw from their IRAs without restric- by an increase in IRA assets. That is because when “traditional” defined benefit (DB) plans that historically Loan Activity in 1988.” —Previous research on employee EBRI Issue Brief no. 218 distributions are increasingly available in DB plans. ing the benefits of life annuities. First, the pricing of amount they match per dollar of employee contribution a decline of 25.0 percent. Defined contribution plans will termination to participants who are entitled to plan Campbell, Sharyn. “Hybrid Retirement Plans: The annuity would provide). Consumers think that the odds plans by adding Sec. 401(l) to the Internal Revenue Code Are Going?” In Dallas Salisbury, ed., termined formula, contribution plans other than 401(k) plans is and, most frequently, all contribu- years.cent, respectively, for Retirement retirees will ) have fiduciary responsibilities for investing trust assets average account balance for those who rolled over their Work Force Individual retirement account 35.2 45.9 30% Chart 8 Rollovers ...................................................................... 24% 7 order to survive in the global economy. The result has Chart 7, Predicted 401(k) Contributions for In general, the model in this analysis uses a combination would lead to higher mortality rates and they would be forever. Research Institute, February 2001) (Employee Benefit Research Institute, February provide 33.7 percent of the retirement wealth for men in benefits. In contrast, because Congress does not require of them dying are in insurance companies’ favor, so they tions (made by both employers and/or employees) are (IRC), it would appear that a significant percentage of Prospects in a Defined Contribution World Retirement Income System Continues to Evolve.” cohort), and the difference males, while the de- Given these trends, it is (Washing- contribution behavior to 401(k) plans has often been in which the employee was participating in a defined tion. Therefore, we also assume that all pre-retirement qualified DC plan participants terminate with the have been the predominant source of employer-provided Thirdly, most annuities offer no protection In addition to the fact that DC plans are becom- annuities discourages their purchase, as the price of and the maximum percentage of compensation they are estimated from self-reported responses to receive that will on behalf of plan participants, and employers directly account balances to an IRA was $68,107, compared with How Much Can Individuals Spend? 20% • This Issue Brief highlights the changes in private pension plan participation for DB Composition of Estimated Retirement Wealth for Males at Social Security Normal Retirement Age Modeling Retirement Income .......................................... 7 Source: Employee Benefit Research Institute, Retirement Income Projection Model. 20% 8 Contact EBRI Publications, (202) 659-0670; fax publication orders to been less employer flexibility to meet unexpected costs, of Form 5500 data from the Department of Labor and Research has found that changes in the work force Selected Persons and Plans remains fairly constant over time ...................................... 14 Hurd, Michael, Lee Lillard, and Constantijn Panis. “An paying for the average mortality rate. Thus, despite 2000). this birth cohort, and 31.9 percent for women. IRAs will DC plan sponsors to insure DC plan benefits, employers will lose. Thus, consumers miss the point that the defined benefit sponsors have retained PIA-offset plans. placed in individual accounts on behalf of each partici- ton, DC: Employee Benefit Research Institute, 1997), EBRI Special Reportreasonable to assume the percent- SR-32/ creases for females are Issue Brief no. 171 (Em- limited by lack of adequate data. This is primarily due to against inflation, which discourages their purchase. contribution plan sponsored by the previous employer. A ing an increasingly important source of future access to IRAs would be prohibited. retirement income. Both of these factors are likely to 14% annuities in the individual market typically diverges willing to match. Employee contribution behavior is have to be managed by the retiree. This makes the risk public survey data. assume all investment risk. be much impact in the short run, given the model’s sponsoring employer, they may avoid current taxation on DB benefit formulas $8,445 for those who cashed out. This finding suggests may be artificially inflating the estimated equity per- According to T Model Description erminated ubscriptions/or and DC plans and provides some possible explanations for these changes. Results Under Baseline Assump ders tions, by Birth Cohort S (202) 775-6312. Subscriptions to EBRI Issue Briefs are included as part of Model Description ....................................................... 8 9% 3 which has made DC plans more appealing because they Chart 8, Composition of Estimated Retirement Since many 65-year-old males will live longer than the self-reported responses to public survey instruments to during the late 1970s to early 1980s (such as changes in (37.2 percent vs. 26.4 percent for Analysis of the Choice to Cash-Out, Maintain, or being already “worse-off” due to bad health, they would the types of matching formulae utilized by sponsors. Yakoboski, Paul J., and Jack L. VanDerhei. “Contribu- Consequently, individuals could be faced with receiving a component was also included in the model to estimate retirement income, DB plans increasingly are adopting their DC plan assets by rolling them over into IRAs. In reduce the amount of life annuity benefits The estimated results of such a modification to that future from the actuarially fair price. This occurs for two decomposed into a series of 1 percent of compensation expand their role the most, reaching 39.9 percent for sponsoring DC plans do not pay benefit insurance purpose of annuities is to provide insurance against the of longevity more central to retirees’ expenditure deci- pant. In comparison, DB plans are typically In order to estimate the offset provided under the plan pp. 55–92. ployee Benefit Research Institute, March 1996). age of defined benefit participants estimated at 2.7 and estimates from directly determine plan benefits owed to participants. assumption with respect to transition provisions. How- 10% that perhaps participants with smaller account balances centages. Preliminary analysis of rebalancing was are presented from the Employee Benefit Research Institute’s (EBRI) Retirement 4% EBRI membership, or as part of a $199 annual subscription to 3% EBRI Notes and EBRI Issue Briefs. Individual copies are Vested Defined 100% 1% Estimating Current and Future Accrued Benefits 0 25 0 Participation offer more cost predictability (for example, employers do The model used in this analysis is based on results from Wealth for Males at Social Security Normal projected average 16.4 years, an individual must make model coverage, participation and initial account balance union participation, firm sizes, and industry sectors) can 35 also lose the most from mandated annuitization. Annuitize Pension Rights at Job Change or Retire- tion Rates and Plan Features: An Analysis of Large Papke, Leslie. “Participation in and Contributions to men and 30.9 percent for women. premiums. Clark, Robert, Gordon P. Goodfellow, Sylvester J. risk of outliving one’s resources. formulae, the model computes the employee’s Average noncontributory, and plan contributions are held in one the 1964 cohort). covered by a cash balance plan 2.9 percent. These formulae are often complicated due to the desire of withdrawals from IRAs. retirement payment that is constantly declining in real DC-like features that may result in fewer life annuity the current system are shown in panel E of tables 1 and fact, a 1997 estimate determined that 22 percent of retirees will receive relative to current retirees, raising reasons. Insurance companies must cover their adminis- intervals, which therefore permits modeling not only the sions. Therefore, they will have to understand that life Investment Returns—Although the model has the U.S. Depart- That is, employers guarantee specific benefit levels to do not appreciate the considerable retirement income undertaken with the EBRI/ICI 401(k) data last year, ever, for males in the 1964 birth cohort, the average available with prepayment for $25 each (for printed copies) or for $7.50 (as an e-mailed electronic file) by calling Income Projection Model that quantify how much the importance of individual 0 and Account Balances ............................................ 8 23 Benefit Plan Participants a four-year time series of administrative data from more expenditure decisions based on a longer time horizon. not assume investment risk under DC plans). These © 2001. information for all defined contribution participants, only explain some of the increasing prevalence of DC Retirement Age Under Baseline Assumptions, The baseline results are Secondly, redistribution would likely occur from poor to ment.” 401(k) Plan Data.” RAND Working Paper. EBRI Issue Brief DRU-1979-DoL. no. 174 (Em- trust on behalf of all participants. Employers offering DB Indexed Monthly Earnings, primary insurance amount, 401(k) Pension Plans.” Schieber, and Drew A. Warwick. “Making the Most of The main result of the model developed in this Journal of Human Resources may have doubled between 1997 sponsors to provide sufficient incentives to non-highly terms. If the assets were held instead of being payments for future retirees. Lump-sum payment 2. As expected, there is a sizable increase in the impor- questions as to whether other sources of retirement trative and sales expenses in the underwriting and marginal incentives to contribute at that interval but expectancies are merely averages, and that wide varia- been designed to generate investment rates of DB plan participants regardless of the plan assets’ proportion of retirement wealth estimated to be derived additions to IRAs in that year were attributable to 20s EBRI or from www.ebri.org. 30s 40s Change of Address: 50s EBRI, 2121 K Street, NW, Suite 600, Washington, DC 20037, (202) 60s that can be amassed over time by preserving even small and while additional analysis remains to be completed, it ment of Labor, 90% account plans is expected to increase because of these changes. This Issue Brief also Results ............................................................................ 11 13 same business pressures have led employers to want than 10 million 401(k) participants and more than Employee by Birth Cohort ......................................................... 14 Cooley, Hubbard, and Waltz (1998) estimated various as well as contribution behavior for non-401(k) defined Trends plans. However, in recent surveys, plan sponsors who based on several assumptions October 1998. 775-9132; fax number, (202) 775-6312; e-mail: Publications Subscriptions@ebri.org. Membership Information: Inquir- compensated employees to contribute in order to comply rich. If all risk classes were treated the same, then there annuitized, increases in nominal interest rates could options are becoming more frequent among DB plans. tance of IRAs at the expense of the other two types of Defined Benefit Plans rollovers, whereas only 2 percent were attributable to income—such as individual account plans (DC plans and ployee Benefit Research Institute, June 1996). —A stochastic job duration marketing of their annuity products, plus some level of also the “option value” that making the contribution at The implications for retirees are major: Issue Brief tion beyond the average is possible. plans must make contributions based on federal funding and covered compensation values for the birth cohort. 30 (Spring 1995): 311–325. 401(k) Plans: Who’s Choosing What and Why.” In is the expected increase in income coming return on a stochastic basis, for purposes of and 2000. In an attempt to model the percentage The significant investment performance. from (non-cash balance) defined benefit plans decreases amounts of DC plan assets. appears that there may be more of a discrepancy be- Age Simulating the Impact of a Prohibition of discusses the risk of outliving one’s assets, since a greater fraction of pension wealth In certain situations, defined benefit plan sponsors have Cash Balance Plans Benefit ies regarding EBRI membership, and/or contributions to EBRI-ERF should be directed to EBRI President Dallas .................................................. 12 benefit plan designs that align retirement benefits with 30,000 plans, as well as a time series of several hundred 80% 16 Chart 9, Composition of Estimated Retirement payout (withdrawal of assets) percentages that would be contribution plans. Asset allocation information is based switched from DB to DC plans said that their primary Hustead, Edwin C. “Trends in Retirement Income Plan would be a redistribution of income from those who die ________. “Large Plan Lump-Sums: Rollovers and Poterba, James M., Steven F. Venti, David A. Wise. • from sources that are not guaranteed for life (i.e., defined rules and regulations in order to maintain the plan’s Most of their non-Social Security retirement income Olivia S. Mitchell, P. Brett Hammond, and Anna M. that may prove to be biased when the potential impact of this shift in partici- with technical nondiscrimination testing. This complex- mitigate some of the effects of inflation. While some This is due in part to the conversion of traditional DB plans. Given that (a) the likelihood of rollover to an IRA model was estimated and applied to each individual in direct IRA contributions individual retirement accounts, or IRAs)—will make up (the remaining 76 percent profit in the premiums they charge. Typically, those who that interval provides for the employee. Chart 7 illus- To illustrate, according to projections from the this Issue Brief the results are obtained from of private wage References A third important way DB and DC plans differ is tween active rebalancers and passive investors among from 26.4 percent to 25.3 percent—a 4.1 percentage is projected to come from “nonguaranteed” sources. the ability to automatically cash out former employees. Source: Paul J. Yakoboski and Jack L. VanDerhei, “Contribution Rates and Plan Features: An Preretirement Cashouts From Defined Salisbury at the above address, (202) 659-0670; e-mail: email@example.com 14 Research Long-Term Asset Allocation for Defined Changes in 32 plan descriptions used to provide a sample of the various sustainable for a different number of years. Across the employee performance and company profits (for on previously published results of the EBRI/ICI 401(k) motivation for doing so was “matching worker character- Wealth for Females at Social Security Normal additional information becomes earlier (who are more likely to be poor) to those who live Administrative Expenses.” Cashouts.” EBRI Issue Brief PRC Working Paper 96-13. no. 184 (Employee contribution qualified (tax-favored) status. Defined Contribution Plans “Lump-Sum Distributions from Retirement Saving will be subject to the retiree’s control regarding the Rapport, Forecasting Retirement Needs and Retire- rather than defined change, the distribution of benefit pation rates sources of ity makes it virtually impossible to appropriately analyze annuity products have been developed around the U.S. plans to cash balance plans. Cash balance plans are DB the model to predict the number of jobs held and the age in the baseline scenario increases with account balance, the difference. choose to purchase annuities live longer than those who trates the predicted employee contributions from the 2000 Social Security Trustees’ Report, a 65-year-old male running it in a deterministic mode. the form in which they generally pay plan benefits. As decrease (table 3). The long-run results for females resulted from investment gains). Although individuals Analysis of Large 401(k) Plan Data,” EBRI Issue Brief, no. 174 (Employee Benefit Research Institute, and salary workers participating in a primary DB plan older employees. The When a worker separates from employment prior to 70% Institute- 24 Contribution Plans ............................................... 15 Contribution Plans and IRAs 29 defined benefit and defined contribution plan provisions example, profit-sharing plans using a discretionary Retirement Age Under Baseline Assumptions, three asset allocations examined, only a 3 percent June 1996). database, istics to plan characteristics” (Quick, 1999). Younger and and employee contribution behavior to available. Therefore, this report Philadelphia, PA: University of Pennsylvania, Pension the employee’s behavior if one is forced to observe either longer (who are more likely to be the wealthy). While government inflation-indexed Treasury Inflation Pro- plans under which employers usually communicate and (b) males tend to have larger account balances than of each job change. Each time the individual starts a new can purchase life annuities with their IRA assets (or Benefit Research Institute, April 1997). This Issue Brief highlights changes in private decreased from 38 percent in 1977 to 22 percent in 1996 do not. Thus, the insurer is faced with an “adverse model as a function of employee demographics and the income). Consequently, retirees will be at a greater risk Initial Account Balances Plans: Receipt and Utilization.” rate of spending, rather than arriving like a regular ment Wealth. Another important way DC and DB plans differ Philadelphia, PA: University of defined benefit plan types for jobs —Previous studies NBER Working from DB to DC on the indicated above, DC plans usually offer lump-sum suggest a 2.6 percent decrease in the importance of For purposes of this sensitivity analysis, an ad trends • Results of the model are compared by gender for cohorts born between 1936 and 1964 normal retirement age, the present value of the accrued Education Modeling Retire- Terminated Vested Defined Benefit Plan Editorial Board: Dallas L. Salisbury, publisher; Steve Blakely, managing editor; Cindy O’Connor, production and distribution. Any applicable to plan participants. In addition, several contribution approach), and to expect employees to by Birth Cohort ......................................................... 15 withdrawal rate was found to have a high probability 401(k) plans is provided by an expansion of a method Asset Holdings more mobile workers are thought not to appreciate this can be mitigated through annuity options (such as Allen, Jr., Everett T., Joseph J. Melone, Jerry S. Research Council, of outliving their resources, and judging from the discus- is in which party directly assumes the investment risk EBRI/ICI 401(k) database have analyzed the average Paper paycheck for life. Pennsylvania Press for the Pension Research Council, . No. W5298. Cambridge, MA: National Bureau explores how sensitive the results taken after the year 2000 was plans and the aggregate plan data or use information on the plan 60% tected Securities, the resulting market has remained benefits as “account balances.” Although cash balance females, cet. par., it is not surprising that the impact of job, the model simulates whether or not it will result in choose annuities as their IRA investment selection), they pension plan participation for DB and DC plans and (the most recent data available, chart 1). During that selection” issue and must price the annuity to account for employer’s matching formulae. Panel A benefits, meaning that the entire amount of accumulated hoc reduction in equity percentage was applied to reflect outlined pension wealth for defined benefit plans other than cash 39 in order to estimate the percentage of retirees’ retirement wealth that will be derived benefits is computed and compared with the statutory and Research The relative incidence of cashouts and rollovers among To calculate stay behinds from SCF, the total number of rollovers from views expressed in this publication and those of the authors should not be ascribed to the officers, trustees, members, or other Participants .......................................................... 16 public surveys based on participants’ self-reported (80 percent) of not depleting one’s assets before 30 years. assume a larger role in helping their employers fund Rosenbloom and Jack L. VanDerhei. Pension Plan- based on both employee demographic information and Table 1, Composition of Estimated Retirement traditional DB plans, which are “back-loaded,” meaning are with respect to: Ippolito, Richard A. refund options) and potentially through tax and subsidy Pension Plans and Employee Perfor- • sion above, people do not seem to be taking advantage of on plan assets and whether that investment risk directly account balances for 401(k) participants by age and of Economic Research, October 1995. Rather than the assets that back up their income 2000. modified by taking a pro-rata rise in the contribution formulae provided by the participant. quite small (Brown 2000). plan “account balances” are notional only, cash balance this change would be larger on females. For females born coverage in a defined benefit plan, a defined contribution provides some possible explanations for them. Next, the longer longevity of those who purchase annuities. While the 401(k) plans used in the previous in table 1 assets in the account are paid out at one time. If the balance. As expected, most of the offsetting increase is may withdraw their IRA assets in essentially any same period, the percentage participating in a primary what is likely to be more of an equilibrium asset alloca- pension plans plus the total number of previous pension plans with assets or above sponsors of the Employee Benefit Research Institute, the EBRI Education and Research Fund, or their staffs. Nothing herein is to from DB plans versus DC plans and IRAs over the next three decades. Under the Chart 6 Fund. threshold for mandatory cashouts at the time. Under the participants taking distributions from their previous 50% Simulating the Impact of a Prohibition of 26 ment Income answers (SCF, the Current Population Survey (CPS), their retirement benefits (for example, cash or deferred 17 rights left behind was determined. The percentage of stay behinds was then Wealth for Males at Social Security Normal The more aggressive investments in equities would ning: Pensions, Profit Sharing, and Other Deferred plan matching provisions. that older workers accrue benefits that are more valu- • Trends in cash balance plans. mance. Chicago, IL: University of Chicago Press, 1997. be construed as an attempt to aid or hinder the adoption of any pending legislation, regulation, or interpretative rule, or as legal, mechanisms by the federal government, the exact plans generally offer benefits in the form of lump-sum plan, both or neither. in 1964, mandatory rollovers are estimated to increase manner they choose, results are presented from the Employee Benefit Re- With the exception of studies based on adminis- Another reason why individuals choose not to including in the form of lump- If coverage in a defined benefit DC plan increased from 7 percent to 23 percent, and the paper (VanDerhei and Copeland, 2001) provided the ________. “Pre-Retirement Cashouts and Foregone number of households with DC plans only have been Clark, Robert L., and Ann A. McDermed. annuities’ protection against this risk. Thus, a possible affects plan benefits. The overwhelming majority of DC tenure (see chart 4). The most recently published results being managed by the sponsor of the pension plan, The second reason why individuals do not reduction across all the final The Choice of provides retiree needs this amount to ensure adequate retirement found in the defined contribution and cash balance plan Company A Contribution Rates by Salary tion by age. The equity reduction for purposes of this indicate All rights model’s baseline assumptions, both males and females are found to have an baseline scenario in this model, it is assumed that employer’s defined contribution plans has been docu- the number of previous pension plans with assets or rights left behind relative Preretirement Cashouts From Defined 11 accounting, actuarial, or other such professional advice. and the Survey of Income and Program Participation compensation DC plans). Retirement Age Under Baseline Assumptions, support a 4 percent withdrawal rate for having a high able as a percentage of compensation than do younger Kusko Andrea L., James M. Poterba, David W. Wilcox. relationship between income and mortality is not known. Compensation Plans. Eighth edition. Homewood, IL: accompanied by an increase in DC plan assets. Private policy topic to reduce this risk is to mandate or encour- plans offer participants a choice of account investment Retirement Saving: Implications for 401(k) Asset which would bear the risk of investment losses, the Pension Plans in a Changing Regulatory Environ- • average categories to rebalance to Long-term asset allocation for trative data, employee contribution behavior is typically plan is predicted, time series information from the annuitize is the thought that they can invest in higher- distributions. the relative importance of IRAs by 45.9 percent, while sum or periodic distributions. search Institute’s (EBRI) Retirement Income Projection percentage of those participating in supplemental DC purchase annuities is the desire for flexibility. The exact matching formulae adopted by the plan sponsor, show that the year-end 1999 average balance ranged the income over the course of his or her retirement, he or she paper was arbitrarily assumed to increase from a 0 per- that reserved. component: The 1964 birth cohort is estimated to experi- to the combined total described previously. Fidelity (2001) provides an 38 40% 60% appreciable drop in the percentage of private retirement income that is attributable Results employers will cash out these amounts whenever pos- mented extensively elsewhere. However, there has 20 59% Contribution Plans and IRAs............................... 16 34 53% (SIPP) probability for assets to last 30 years. Thus, what most Richard D. Irwin, Inc., 1997. ) were used to model participation, wages, and 20 workers. Generally, back-loaded DB plans provide the analysis of their experience with respect to participant choices after by Birth Cohort ......................................................... defined contribution plans. 12 “Employee Decisions with Respect to 401(k) Plans: Max Match (5%) trusteed DC plan assets first surpassed private trusteed age (give preference to) the annuitization of all defined options, and plan participants directly assume all from $4,479 for participants in their 20s with less than Accumulation.” individual will have to self-manage the assets or select ment. Washington, DC: AEI Press, 1990. NBER Working Paper 100 percent. . No. W7314. assumed to be a function of employee demographic data Bureau of Labor Statistics (BLS) is used to predict return assets than what an annuity can provide. the importance of both defined benefit plans and the Model that quantify how much the importance of indi- the vast majority of the 30,000 plans in the EBRI/ICI composi- must manage (e.g., invest and spend) the amount in a ence a 2.9 percent increase in retirement wealth in these plans gradually increased from 10 percent to 16 percent. The Survey of Income and Program Participation (SIPP) is a longitudinal future retirement income will depend less on “guaran- cent reduction for 20-year-olds to a 10 percent reduction EBRI Issue Brief is registered in the U.S. Patent and Trademark Office. ISSN: 0887-137X 0887-137X/90 $ .50+.50 to defined benefit plans (other than cash balance plans). In addition, results show sible, and that the assets will not be retained in the been a lack of data with respect to the long-run behavior termination. Based on 1999 job terminations, they found that the percentage Discussion ...................................................................... 17 50% 47% initial account balance information. 30% Plan Max (9%) Table 2, Composition of Estimated Retirement survey conducted by the U.S. Census Bureau to collect data on income sources people would consider to be rather low withdrawal rates American Council of Life Insurers, Life Insurers Fact majority of plan benefits in the final several years before • Cash-out behavior for defined Evidence From Individual-Level Data.” NBER Work- Regulator and perhaps an employee’s estimate of the employer- whether it will be: However, individuals can purchase annuities linked to combination of defined contribution and cash balance vidual account plans is expected to increase because of y Environment Most analysis Two prevailing Participant-Directed Retirement Plan Data Collection defined benefit plans assets in 1997 and continued to contribution plan and IRA assets or a portion of those Cooley, Philip L., Carl M. Hubbard, and Daniel T. Walz. investment risk. DC plan benefits are determined by Cambridge, MA: National Bureau of Economic Re- someone to do so. Not surprisingly, higher DC plan participation The results for panel B in tion of manner that ensures that outcome. Otherwise, the categories for both males and females. The impact on the teed” forms of benefits (such as DB plan annuities and for 65-year-olds. Rollovers a clear increase in the income retirees will receive that will have to be managed by 43% of assets that remained in the plan until the end of the year varied from 33.7 retirement system. However, there does not appear to be of individuals who—at least initially—leave their 42% and amounts, labor force participation, program eligibility and participation, 402(g) How Much Can Individuals Spend? ......................... 18 41 33 The model attempts to estimate the balance of Wealth for Females at Social Security Normal would be necessary to protect an individual against retirement. In contrast, most DC plan benefits are less ing Paper. Book. Washington, DC: American Council of Life No. 4635. Cambridge, MA: National and public • grow steadily relative to DB plan assets through 1999 This section closely follows Brown (2000). assets to achieve an income stream above some level percent for participants with account balances of less than $10,000 to 68.7 search, August 1999. Rather than having the sponsor of the pension plan “Retirement Savings: Choosing a Withdrawal Rate tables 1 and 2 are generated with benefit terminated vested matching rate (Bassett et al., 1998) or a proxy based on • equity returns that allow them to reap the benefits if plans would decrease by 20.5 percent. Their male A non-integrated (with Social Security) career-average reasons behind rates have led to an increase in the percentage of house- Project do not contain that information, and due to strict estimated retirement wealth for males at retiree runs the risk of outliving his or her lump-sum 37% Social Security) and more on DC plan benefits—which The results of this sensitivity analysis are shown IRA percentages is estimated to be de minimis for both 40% and general demographic characteristics. Previously, households were the retiree. This makes the risk of longevity more central to retirees’ expenditure a source of time series information on the employer’s 20% account balances with the previous employer when they Why the Shift The cost of complying with new and changing require- Annuities ................................................................... 18 percent for those with account balances in excess of $200,000. However, as outliving his or her assets. any defined contribution plan and/or IRA (whether Before retirees need to make decisions about how to 42 followed for two and one-half years, but starting in 1996 households were age-sensitive. Thus, benefits payable upon job termina- Retirement Age Under Baseline Assumptions, Did you read this as a pass-along? Stay ahead of employee benefit participants. Bureau of Economic Research, February 1994. Insurance, 2000. discussion of Quick, Carol. “An Overview of Cash Balances.” (chart 3). Furthermore, assets held in IRAs, which are such as the poverty level. However, there are benefits 25 bear the risk of the retiree living to 100, the retiree That Is Sustainable.” the same assumptions as panel A, AAII Journal (February 1998): EBRI Form 5500 data (Papke, 1995). However, as shown in higher returns occur, while still having the protection counterparts are estimated to experience an increase of plan. 31% the increasing confidentiality standards no information on the plan Social Security normal retirement age under genders. Could we send a friend or colleague a complimentary There are also annuity-certain contracts that make payments for a fixed in panel C of tables 1 and 2. As expected, the decreased This component was based on results reported in Sabelhaus (2000) which 37 decisions. 28% 28% they point out, this may not be sufficient time to observe the long-term propensity to avail themselves of this option. There- change jobs. Therefore, we combine industry data with 27% followed for four years. ments has been higher for DB plans than for DC plans. Conclusion 30% ...................................................................... 20 issues with your own subscription to EBRI Issue Briefs for only $49/ 15 number of periods regardless of the survival of the insured. Thus, no combined survey data on IRA balances with individual tax return data on by Birth Cohort ......................................................... 13 funded through regular contributions or rollovers) of the Bajtelsmit, Vickie L., and Jack VanDerhei. “Risk Aver- spend their IRA assets, they need to make decisions in Laitner, John, and F. Thomas Juster. “New Evidence on Kusko et al. (1994), a significant percentage of the 10% • against outliving their resources. If an individual Conclusion 35.2 percent in the importance of IRAs, while the other An integrated career-average plan. copy of EBRI Issue Brief? 24% retirement role of DC plans sponsor’s identity was included. However, the database also individual account plans, and costs to such a move. incidence, especially for those that terminate late in the year. Hurd, Lillard Notes, will either need to purchase an annuity that transfers 16–21. no. 7 (Employee Benefit Research Institute, the baseline assumptions, by birth cohort. with the exception of the new surpassed both DC and 1 equity concentration results in a long-term decrease for 4 The post-World War II demographic wave of children born between 23% 16 Lump-sum benefits provide the beneficiary with his or her total accrued fore, this sensitivity analysis relaxes that assumption See Sabelhaus (1999). 21% 21 insurance is offered against outliving one’s resources. IRA flows to study IRA accumulation and withdrawal patterns across cohorts. year electronically e-mailed to you or $99/year printed and mailed. 21 For example, Hustead (1996) estimated the cumulative References to DC Plans? ...................................................................... 21 Future versions of the model will include a subroutine for early retirement and Panis (1998) also investigate cashout behavior among defined contribu- Annuities individual at Social Security normal retirement age. In Table 3, Percentage Change in Composition of 18% 36 Altruism: A Study of TIAA-CREF Retirees.” order to accumulate those assets in the first place. 1946–1964, consisting of about 77 million people. sion and Pension Investment Choices.” In Michael S. plans over the American DB trusteed plan assets in 1999. That year, IRA assets Copeland, Craig, and Jack VanDerhei. “Personal Retire- plan benefit generally in a single payment. July 1999): 1–8. that risk to an insurance company or carefully man- The mandating of annuities would bring all risk employee contribution behavior appears to be deter- • attempted to try to self-annuitize by using equity market two categories each decrease by 23.4 percent. A five-year final-average plan without integration. and the trend assumption for cash balance 20% both defined contribution plans and IRAs. Cash Balance Plans Overall, 16% 6 • The implications of these model results for retirees are significant. First, individu- Long-Term Asset Allocation for Defined For more information about subscriptions, visit our Web site at and assumes that departing employees are not cashed itself. However, the Pension Benefit Guaranty Corporation insures defined 43 behavior. tion plan participants; however, it is subject to the same limitations 17 Employer contributions may be subject to vesting rules, such that partici- 26 10 Subject to penalties for early withdrawal and minimum distribution 0% 13% 13% This is different from the tremendous growth in variable annuities in the Thus the current version of the model ignores the possibility that an costs of regulatory changes promulgated from 1981–1996 Send an issue to See Copeland and VanDerhei (2000). The Survey of Consumer Finances addition, it estimates the accrued benefits earned and Estimated Retirement Wealth at Social Security Economic Review. Specifically, employees who leave a job must decide to 2 Gordon, Olivia S. Mitchell, and Marc M. Twinney, Vol. 86 (1996): 893–908. last 20 years Sabelhaus, John. “Modeling IRA Accumulation and equaled $2.47 trillion, compared with $2.45 trillion and classes into the market, and not just those who think 5 age his or her individual rate of spending to avoid ment Plans: An Analysis of the Survey of Consumer plans. There does not appear to mined by plan-specific provisions. For example, in chart • diversification, even with a reasonable fraction of wealth A three-year final-average plan without integration. Poterba, Venti, and Wise (1999) estimated the toward incorpo- Based on current intermediate projections, after the trust fund is depleted in benefit accruals up to a limit, thereby reducing that risk. Poor investment there was estimated to be a 0.8 percent decrease in the mentioned in footnote 37 above. Technically, most private qualified defined contribution plans are either 1 pants do not have full legal right to employer contributions made on their 37 www.ebri.org or complete the form below and return it to EBRI. als—rather than the pension plan sponsor—increasingly will have to manage their requirements that apply beginning at age 70 /2. 1936 1938 1940 1942 1944 1946 1948 195022 past decade. However, variable annuities are deferred annuities. Under a Hurd, Lillard, and Panis (1998) find that “among DB plan holders with a employer will adopt a new retirement plan—or change an existing plan—prior 1952 1954 1956 1958 1960 1962 1964 Even using the small withdrawal rates suggested above, out and remain in terminated vested status until normal (SCF) is a triennial survey, conducted by the Federal Reserve Board, which Contribution Plans 10% An alternative method of comparison would leave the defined benefit for employers offering DB versus 401(k) (DC) plans. He 2037 current-law benefits would need to be reduced by more than 20 percent performance also may indirectly affect plan benefits if it results in a money purchase or profit sharing (Sec. 401(k) plans are of this type). Under Normal Retirement Age for Various Scenarios 32 The Bureau of Labor Statistics assumed to be retained by defined benefit plan partici- behalf until they have reached a certain minimum number of years of service. roll over their assets into an IRA from a qualified plan or Organization eds., Positioning Pensions for the Twenty-First Cen- McCarthy, Mike, and Liz McWhirter. “Are Employees 5, the percentage of employees contributing up to either has focused upon the growth of defined contribution • rating DC plan features in DB plans are shifts in the reduction, Milevsky and Robinson (2000) found that impact of preretirement withdrawals on 401(k) asset A five-year final-average plan with covered compensa- lump-sum distribution option, 48 percent started collecting benefits, 31 deferred annuity, individual contract owners are in the process of accumulat- $2.21 trillion for trusteed DC and DB plans assets, they are going to live a long time. Thus, annuities could 40 to the employee’s job separation and/or retirement. Withdrawals.” outliving the assets. Finances.” The model assumed a CPI growth rate of 3.50 percent, a EBRI Issue Brief National Tax Journal. no. 223 (Employee Vol. 53, no. 4 collects comprehensive, nationally representative data on the wealth of defined contribution plan percentage component for 30 retirement assets and bear the risk of investment losses. Second, since most retirees’ accruals as annual benefits and “annuitize” the defined contribution and IRA 18 An estimate of the true impact of such a proposal would need to account 402(g) denotes the IRC section that limits employee deferrals to a specific the risk of outliving one’s resources is not eliminated retirement age when they receive their deferred retire- See McCarthy and McWhirter (2000). across-the-board. 0 curtailment of future benefit accruals or affect the employer’s ability to provide the former, the plan sponsor typically commits to a fixed percentage of 0 0 0 0 0 0 estimated that the cost of administering a DB plan in 12 percent expected to draw benefits in the future, 4 percent took an LSD and ing assets, not receiving retirement income that is longevity-insured. Name real rate of return for stocks of 6.98 percent, and a real rate of (U.S. Department of Labor, 1999) pants, and converts this amount to a present value at 7 American households. It collects data on households’ total liabilities and Relative to the Baseline Assumptions: 1964 Defined Contribution + Cash Balance Defined Benefit plans and the decline of defined benefit plans. Little Missing the Big Picture?” to leave their assets in their existing DC plan, as op- tury. Philadelphia, PA: University of Pennsylvania 0 Benefits Quarterly Individual Retirement Account (First balances. However, this presents additional complications in modeling be priced using the overall average mortality characteris- for, inter alia, the potential impact that such a reform would have on: (1) the (December 2000): 865–875. Benefit Research Institute, July 2000). the IRC Sec. 402(g) limit, the maximum amount of Forfeitures arise when employers terminate employment and leave work force and changes in the business environment. between 17 percent and 32.5 percent of males age 65 accumulations and concluded that it reduces average Address tion as the integration level. amount per year. Unlike the other two variables, this value is constant across respectively. 27 males in the 1964 birth cohort and a 2.1 percent decrease ad hoc benefit increases to retirees. Some employers offer ad hoc benefit compensation each year. For a profit-sharing plan, plan contributions may be For additional detail on integrated defined benefit plans, see Chapter 4 of To the extent that some defined contribution plan non-Social Security retirement income will be distributed as a lump sum or in completely. The insurance vehicle for ensuring against ment benefits. rolled it over into an IRA, and only 16 percent cashed out their pension Furthermore, assets in variable annuities are not required to be converted to a 19 return for bonds of 3.00 percent. In addition, 1 percent is 3 assets, including pension wealth. Rather than rolling over or cashing out defined contribution plan assets, Social Security pays life annuity benefits, and most traditional, employer- 10-20 20-30 30-40 40-50 50-60 60-75 75-100 > 100 employee purchasing behavior. relative likelihood that employers would want to continue sponsoring such a 34 1981 was approximately140 percent higher than the cost nonvested benefits in the plan. Nonvested portions of any terminating all plans for any particular year. reports that 6 percent of full-time normal retirement age for comparison of relative magni- increases to offset the effects of inflation on the value of DB plan benefits. made on a discretionary basis by the plan sponsor, but how these contribu- Allen, Melone, Rosenbloom and VanDerhei (1997). Birth Cohort, by Gender This implicitly assumes that transition ........................................... 16 analysis has been available of the long-term implications Quarter 2000): 25–31. posed to “cashing out” their benefits and spending them. Press, 1997. ________.“Projecting IRA Balances and Withdrawals.” Organization tics, thereby bringing prices closer to their actuarially Elliot, Kenneth R., and James H Moore, jr. “Cash Most Americans have not faced these challenges compensation matched or the plan maximum is shown as • 401(k) assets at age 65 by about 5 percent. Comparing The increasing role of DC plans can further be explained A three-year final-average plan with covered compen- rights.” However this information is limited to a relatively older population life annuity and evidence currently suggests that not many are being converted subtracted from each of the stock and bond real rates of return City/State/ZIP 41 participants do not actively rebalance their asset alloca- periodic payements (from a defined contribution plan or IRA) rather than as a participants may leave their assets in their previous employer’s plan. While sponsored defined benefit plans pay benefits in the form of life annuities. In plan and, if they did, how their contribution formulae may need to be outliving one’s resources is a life annuity. employees’ accounts may be used to reduce employer contributions or may be The results are shown in panel D of tables 1 and This type of 11 Salary ($thousands) 23 tions are allocated among individual employee accounts must be based on a provisions allow everyone in a traditional of administering a 401(k) plan. However, by 1996, DB 18 See Quick (1999). 31 (the 1931–1941 birth cohort) and the authors confine their analysis to reports (Brown and Warshawsky 2000). Holden and VanDerhei are currently analyzing similar activity exclusively to reflect administrative costs. employees in medium and large tudes of those benefits that are typically thought of as Bassett, William F., Michael J. Fleming, and Anthony P. In a study of Hewitt 401(k) data, 57 percent of partici- 9 Clark, Goodfellow, Schieber and Warwick (2000) also use administrative 28 Milevsky, Moshe Arye, and Chris Robinson. “Self- 45 the assets or balances remain in the plan, they are still accumulating a function of age for one of the plans studied by of this change on the composition and levels of future the results from this scenario (for the combination of by the fact that DB plan requirements under the Em- general, life annuity benefits are payments made on a periodic basis, often sation as the integration level. in the past, nor has the nation focused on financial Seventy-six percent of full-time workers participating in a DB plan in a fair level. In fact, Mitchell et al. (1999) found that restructured, and (2) how employees (especially those designated as EBRI Notes, BLS information was utilized to code the distribution of generosity Balance Pension Plans: The New Wave.” no. 5 (Employee Benefit Research Compensa- reallocated among the remaining defined contribution plan participants. There remains a debate on whether individuals do behave in a manner Address 35 tions to reflect abnormally high recent experience in the regular paycheck for life (from a defined benefit plan), retirees will need either to specified, predetermined formula meeting certain requirements if the plan is to defined benefit plan that is converted to cash Your Name VanDerhei, Holden, and Quick (2000). annuity pays regular payments for the length of one or 2. As expected, this new assumption increases the of plan dispositions for respondents who left their jobs between waves 1 and 2 within the 401(k) population by using the EBRI/ICI 401(k) database. 15 data but only investigate the match rate, not the maximum amount of 12 investment gains (or losses). Participants who choose this option may monthly, for the life of the beneficiary. In addition, life annuities frequently Source: Employee Benefit Research Institute, Retirement Income Projection Model. 30 44 medium or large establishment were not offered a lump-sum distribution in Individual retirement accounts (IRAs) can be established by individual parameters for flat benefit plans. nonhighly compensated employees under IRC Sec. 414(q)) would decrease 33 private establishments had a “guaranteed” (usually by an implicit employer annuity) consistent with a bequest motive. For example, see Laitner and Juster (1986) This follows Gale, Papke, and VanDerhei (1999). retirement income. Annuitization and Ruin in Retirement.” pants who removed their assets from a previous Source: Paul J. Yakoboski and Jack L. VanDerhei, “Contribution Rates and Plan Features: An Analysis of Rodrigues. “How Workers Use 401(k) Plans: the North education and financial literacy in a sustained and Mitchell and Moore (1998) found that Social Security and private pensions annuity payments could increase up to 10 percent if qualify for tax-favored status. Institute, May 1999): 1–4. tion and Working Conditions Defined as the value of account balances from defined balance to remain in the current plan until (Summer 2000): 3–11. 8 Yakoboski and VanDerhei (1996). • ployee Retirement Income Security Act of 1974 (ERISA) IRAs, defined contribution, and cash balance plans) with A five-year final-average plan with a PIA (primary Chart 6 shows 42 of the HRS. Investment risk affects defined benefit plan participants’ benefits indirectly. equity markets, it appears that using the current asset purchase an annuity from an insurance company or carefully manage their indi- compensation matched and/or the maximum amount of compensation City/State/ZIP 36 withdraw those assets at a later date by taking periodic installments, a lump- more persons’ lives. importance of defined benefit plans relative to the have additional provisions for benefit payments to the beneficiary’s spouse or This publication is available for purchase online. Visit www.ebri.org/publications or call (202) 659-0670. Currently, the market for life workers or can be offered as the funding vehicle under simplified employee 24 1997 (U.S. Department of Labor, 1998). This is down from 85 percent in 1995 their participation and/or contributions if they were not able to withdrawal for supportive evidence of behavior consistent with a bequest motive and IRAs are assumed to have the same asset allocation as defined contribution EBRI Issue Brief, no. 174 (Employee Benefit Research Institute, June 1996). make up approximately two-thirds of the wealth of households nearing See Holden and VanDerhei (2001) on the EBRI/ICI Participant-Directed contribution plans, IRAs and cash balance plans and the they change jobs. This assumption will be Large 401(k) Plan Data,” 27 29 “cash account” benefit formula. 13 versus those that are typically perceived as a lump sum employer’s plan cashed out their assets, 6 percent rolled Holden and VanDerhei (2001) . Poor investment returns, for example, may affect the funding status of the plan American Actuarial Journal. Participation, Contribution, and Withdrawal Deci- This original work by EBRI provides a clear Vol. 4, no. 4 (October allowed by the plan. U.S. Department of Labor. Bureau of Labor Statistics. comprehensive way in the past. While there has been individuals were forced to annuitize their retirement Fidelity Investments. See Clark and McDermed (1990) and Gustman and Steinmeier (1992). Building Futures: Opportunities similar information as a function of salary. It would the baseline assumptions, we find a much larger impact and various tax laws have grown in number since the insurance amount) offset. sum cashout, or by rolling them over to another qualified plan or an IRA. 14 widow/widower. pension plans. (U.S. Department of Labor, 1999). the funds prior to retirement age. Brown (1999) for evidence against. 38 plans in the model. allocation of equities by age for long-run asset allocation vidual rate of spending in order to avoid outliving their assets. See Ippolito (1997) for additional detail. retirement. Retirement Plan Data Collection Project. See Yakoboski (1997) for an example. present value of accruals from other defined benefit plans. relaxed in a future version of the model. annuities other than Social Security and private pen- baseline assumptions. For males born in 1964, there is 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 232 • April 2001 • © 2001. EBRI 6 4 20 10 16 18 14 8 12 24 22 2 April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief April 2001 • EBRI Issue Brief 13 11 19 23 17 21 15 7 3 1 9 5 Table 1 Percentage of P Composition of Estimated Retirement Wealth for Males at Social Security Normal Retirement Age articipants Table 2 Percentage of Participants Percentage of Compensation Composition of Estimated Retirement Wealth for Females at Social Security Normal Retirement Age under Baseline Assumptions, by Birth Cohort $ Trillions Under Baseline Assumptions, by Birth Cohort Issue Brief Type of Plan 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 Type of Plan 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 Panel A: Baseline Assumptions Issue Brief Defined benefit 39.0% 38.5% 38.4% 38.2% 37.3% 36.2% 34.1% 32.7% 32.7% 32.8% 31.3% 30.3% 28.6% 27.0% 26.4% Panel A: Baseline Assumptions Defined contribution + cash balance 33.2 32.3 32.3 32.8 33.3 35.1 37.1 36.5 35.6 34.8 33.9 33.5 34.2 33.6 33.7 Defined benefit 49.7% 48.9% 47.1% 45.7% 44.7% 44.6% 43.2% 42.7% 42.7% 41.1% 40.5% 39.3% 39.9% 37.9% 37.2% Individual retirement account 27.8 29.1 29.2 29.0 29.4 28.7 28.9 30.8 31.6 32.4 34.8 36.2 37.2 39.3 39.9 Defined contribution + cash balance 32.5 32.0 32.6 34.0 34.9 34.3 34.1 33.6 33.3 33.8 32.9 32.1 32.0 32.6 31.9 Panel B: Cash Balance Plans Double Individual retirement account 17.8 19.1 20.3 20.3 20.4 21.1 22.7 23.7 24.0 25.1 26.6 28.7 28.1 29.4 30.9 Defined benefit 39.4 38.6 38.6 38.2 37.0 35.9 33.8 32.1 32.1 32.1 30.4 29.3 27.6 25.9 25.3 Panel B: Cash Balance Plans Double Defined contribution + cash balance 33.0 32.3 32.3 32.8 33.5 35.3 37.3 37.0 36.1 35.4 34.7 34.5 35.1 34.6 34.7 Defined benefit 49.7 48.9 47.0 45.7 44.6 44.4 43.2 42.2 41.8 40.0 39.2 37.9 38.6 36.9 36.3 Individual retirement account 27.6 29.1 29.1 29.0 29.5 28.8 28.9 30.9 31.8 32.5 34.9 36.2 37.3 39.5 40.0 Defined contribution + cash balance 32.5 32.0 32.6 34.0 35.0 34.4 34.2 33.9 34.0 34.6 34.0 33.3 33.2 33.6 32.8 Panel C: Assuming a Less Aggressive Individual retirement account 17.8 19.1 20.3 20.3 20.5 21.1 22.6 23.8 24.2 25.3 26.8 28.8 28.1 29.5 30.9 Long-Term Asset Allocation Panel C: Assuming a Less Aggressive Defined benefit 39.2 38.8 38.8 38.8 38.0 37.0 35.1 33.8 34.0 34.1 32.7 31.8 30.1 28.6 27.9 Long-Term Asset Allocation Defined contribution + cash balance 33.1 32.3 32.2 32.5 33.0 34.8 36.6 36.1 35.3 34.5 33.5 33.2 33.9 33.3 33.4 Defined benefit 49.8 49.2 47.5 46.3 45.5 45.4 44.2 43.9 44.0 42.7 42.2 40.9 41.6 39.7 39.1 Individual retirement account 27.7 28.9 29.0 28.7 29.0 28.2 28.3 30.0 30.8 31.5 33.8 35.1 36.0 38.1 38.6 Defined contribution + cash balance 32.4 31.8 32.4 33.6 34.4 33.9 33.8 33.1 32.6 32.9 32.0 31.4 31.3 32.0 31.2 Panel D: Assuming Terminated Vested Defined Benefit Participants Individual retirement account 17.7 19.0 20.1 20.1 20.2 20.7 22.0 23.0 23.3 24.5 25.8 27.7 27.0 28.3 29.8 Are Not Automatically Cashed Out Panel D: Assuming Terminated Vested Defined Benefit Participants Defined benefit 41.0 40.5 40.3 40.0 39.0 37.8 35.7 34.3 34.3 34.3 32.9 31.9 30.0 28.5 27.7 Are Not Automatically Cashed Out Defined contribution + cash balance 32.1 31.3 31.4 31.8 32.4 34.2 36.1 35.7 34.8 34.0 33.2 32.8 33.5 33.0 33.1 Defined benefit 53.2 52.2 50.1 48.5 47.6 47.3 45.9 45.2 45.1 43.4 42.8 41.5 41.8 39.7 39.0 Individual retirement account 26.9 28.2 28.3 28.2 28.6 28.0 28.1 30.0 30.9 31.6 34.0 35.3 36.4 38.5 39.2 Defined contribution + cash balance 30.3 29.9 30.8 32.2 33.1 32.6 32.6 32.1 31.9 32.5 31.7 30.9 31.0 31.8 31.0 Panel E: Assuming All Defined Contribution Plan Account Balances That Individual retirement account 16.5 17.8 19.1 19.3 19.4 20.1 21.6 22.7 23.0 24.1 25.5 27.6 27.1 28.6 30.0 Are Not Retained in a Defined Contribution Plan on Job Termination Panel E: Assuming All Defined Contribution Plan Account Balances That Must be Rolled Over to an Individual Retirement Account (IRA) and Are Not Retained in a Defined Contribution Plan on Job Termination Preretirement Withdrawals from an IRA Are Prohibited Must be Rolled Over to an Individual Retirement Account (IRA) and Defined benefit 38 37 36 35 34 32 30 29 28 28 26 25 23 21 20 Preretirement Withdrawals from an IRA Are Prohibited Defined contribution + cash balance 33 31 30 30 30 32 33 32 31 29 28 27 27 26 26 Defined benefit 53.0 51.8 49.4 47.4 46.3 45.7 43.9 43.1 42.7 40.6 40.0 38.4 38.5 36.4 35.7 Individual retirement account 29 32 33 35 36 36 37 40 41 43 46 48 50 53 54 Defined contribution + cash balance 30.2 29.7 30.4 31.5 32.2 31.5 31.2 30.7 30.2 30.4 29.7 28.6 28.6 29.2 28.4 Individual retirement account 16.8 18.5 20.1 21.1 21.6 22.8 24.9 26.2 27.1 29.0 30.3 33.0 32.9 34.4 35.8 Source: Employee Benefit Research Institute, Retirement Income Projection Model. Source: Employee Benefit Research Institute, Retirement Income Projection Model.