Printer Friendly

Financing retirement: the private sector; The shift from defined benefit pensions to personal retirement accounts is likely to benefit retirees.

Since 1980, there has been a rapid shift from employer-based, defined benefit pensions to employee-controlled personal retirement accounts. This paper documents the shift and explores the conventional wisdom that this shift increases risk for retirees and will result in lower accumulation of retirement assets. In particular, it focuses on personal retirement accounts and considers the options available for retirees to contain risk and assess the likely outcomes over alternative options, including life cycle allocations. It concludes that personal retirement accounts are likely to lead to higher retirement accumulations that are also less risky than would be the case under defined benefit plans.


Over the past two and a half decades there has been a fundamental change in saving for retirement in the United States, with a rapid shift from saving through employer-managed defined benefit pensions to defined contribution saving plans that are largely controlled by employees. In 1980, 92 percent of private retirement saving contributions went to employer-based plans; 64 percent of these contributions were to defined benefit plans. By 1999, about 88 percent of private contributions were to plans in which individuals decide how much to contribute to the plan, how to invest plan assets, and how and when to withdraw money from the plan. The most important type of personal retirement account is a 401(k), but personal plans also include 403(b), 457(b), traditional and Roth Individual Retirement Accounts (IRAs), and other plans. (1) In the subsequent discussion, "401(k)" includes all of these personal retirement plans.

I will first review the spread and current role of 401(k) plans and then consider the future role of 401(k) plans in retirement saving. I will then comment on the risk of such plans and how retirees may evaluate the risk of alternative investment options. There has been a great deal of comment in the press and elsewhere about the risk of personal retirement plans, perhaps especially following the Enron collapse. The comments often emphasize the market risk of personal plans. Risk compared to what? The implicit assumption is often that employees bear the risk under personal accounts, while under defined benefit plans the risk is born by employers, and there is no risk to employees. However, the bankruptcy of United Airlines and other airlines, for example, makes clear that employees can also be at risk under defined benefit plans. More important, however, the job change risk to employees under defined benefit plans is very substantial. I will not discuss further in this paper the risk under defined benefit plans. (2) Rather, I will consider evidence on how retirees may value uncertain wealth accumulations under alternative investment allocations options in 401(k) plans. I will also consider how risk in such plans might be reduced by life cycle funds and other means and how much employees might gain from such investment strategies.

The spread of 401(k) plans

Figure 1 is a cohort representation of the rapid increase in eligibility for a 401(k) plan since 1984. The figure shows data for nine cohorts, five years apart, denoted by the age of the cohort in 1984. These cohort data were calculated from several waves of the Survey of Income and Program Participation (SIPP). Data are available for seven years noted by the markers--1984, 1987, 1991, 1993, 1995, 1998, and 2003. Each line shows the increase in family eligibility as the cohort ages. Consider cohort C25 that was 25 years old in 1984. In 1984, about six percent of the C25 cohort families were eligible for a 401(k) plan. (3) By 1987, when the cohort was age 28, the percent that was eligible had risen to about 17 percent. By 2003, when the cohort was age 44, eligibility was almost 70 percent. The most important feature of the figure is the increase in eligibility over time for families of a given age. For example, the dashed vertical line highlights the increase in the eligibility of families in cohorts that attained age 40 in successively later years. Cohort C40 was 40 years old in 1984 and about 18 percent of the C40 cohort was eligible for a 401(k) at age 40. Cohort C35 attained age 40 in 1989 and about 34 percent of the C35 cohort was eligible for a 401(k) plan at age 40. The C25 cohort was age 40 in 1999 and about 65 percent of the C25 cohort was eligible for a 401(k) plan at age 40. Similar increases in eligibility are evident at other ages.

Compare the cohort that was age 30 in 1984 with the cohort that was age 30 in 2003. Only 11.8 percent of those aged 30 in 1984 were eligible for a 401(k), while 61.3 percent of those aged 30 in 2003 were eligible. At age 45, 19.5 percent of the cohort that reached that age in 1984 were eligible for a 401(k); while 69.9 percent of the cohort that attained that age in 2003 were eligible.

Figure 2 shows 401(k) participation data for the same 9 cohorts represented in Figure 1. As with eligibility, the increase in participation was dramatic. At age 30: only 5.5 percent of those who reached age 30 in 1984 participated in a 401(k) plan, compared to 43.9 percent of those who reached age 30 in 2003. Only 11.7 percent of those who reached age 45 in 1984 participated in a 401(k) plan, but 56.6 percent of those who reached age 45 in 2003 participated.

The participation rate of those who were eligible for a 401(k) plan also increased. For example, of those who reached age 45 in 1984, 60.0 percent of those who were eligible participated. In contrast, 81.0 percent of those who reached age 45 in 2003 participated if they were eligible.

The rapid increase in eligibility and participation shown in the cohort figures has led to very large increases in contributions to private pension plans.

Figure 3 shows contributions to all private plans between 1975 and 1999 (the last year for which these data are available). Contributions to defined benefit pension plans varied a great deal over the period but were about the same in 1999 as in 1975. Contribution to 401(k) plans began in 1982 and by 1999 had reached $152 billion. Indeed, the increase in total private pension contributions between 1982 and 1999 was essentially accounted for by the rapid increase in contributions to 401(k) plans. Contributions to traditional employer-provided defined contribution plans (DC in the Figure 3) changed very little over this period. Contributions to IRA plans were substantial between 1982 and 1986, but then declined precipitously after the tax advantage of IRAs was reduced for some wage earners. As the figure suggests, I suspect that contributions to personal retirement accounts would likely have been greater today had the tax advantage of IRAs not changed. After 1982, contributions to IRA and Keogh plans combined changed little. (4)

Future Contributions to 401(k) Plans

James Poterba, Steven Venti, and I (2005) have made estimates of contributions to 401(k) plans through 2040. The projections depend first on future demographic trends and future earnings. In addition, they depend on assumptions about participation rates, contribution rates, investment allocation of contributions, the rate of return on contributions, and assumptions about job separation and cash-outs before retirement. Key assumptions behind the projections are these: (1) contributions are invested 40 percent in bonds and 60 percent in stocks, (2) the rates of return on stocks and bonds are assumed to be the same as the average returns over the 1926 to 2003 period, and (3) the contribution rate (including both employee and employer contributions) is assumed to be nine percent, about the same as over the past 10 or 15 years. (5)

Figure 4 shows projected 401(k) assets at retirement (assumed to be age 65) by year of retirement, from 1982 to 2040. The figure shows average assets for all persons retiring in a year, including those who do not have a 401(k) plan. The projected assets at retirement are $24,447 in 2000, $124,984 in 2020, and $408,203 in 2040.

The projections, especially in the distant future are subject to substantial error. For example, suppose the rate of return on stocks was 300 basis points lower than the average rate of return between 1926 and 2003. In this case, the projected average assets at retirement would follow the dashed path, attaining $241,507 in 2040. In either case, the projected increase is very substantial.

For comparison, Johnson, Burman, and Kobes (2004), for example, estimate that the mean present value of employer sponsored pension income for persons 65 to 69 in 2000 was $50,203. This present value calculation is comparable to our projected 401(k) wealth. At age 65, projected 401(k) wealth reaches this level in about 2008 and thereafter increases to much higher levels (in 2000 dollars).

One reason that projected 401(k) wealth is so large relative to defined benefit plan wealth is that the average contribution per participant to 401(k) plans has been about twice as large as the average contribution per participant to defined benefit plans. (6) There are additional reasons why the projections may underestimate accumulation of retirement wealth in 401(k) plans. Because of increases in 401(k) contribution limits, contributions per participant are likely to increase. Perhaps more important, 401(k) plans do not have the early retirement incentives inherent in most defined benefit plans. Thus, employees covered by 401(k) plans are likely to work longer than employees covered by defined benefit plans and thus are likely to have more wealth when they leave the labor force than the projections suggest.

Are 401(k) Plans Risky? What is Best for Retirees?

Compared to what? As I noted above, commentators commonly emphasize the market risk of 401(k) plans, taking for granted that risk under defined benefit plans is born entirely by employers. But the job-change risk inherent in defined benefit plans is substantial. Indeed Samwick and Skinner (2004) find that the job-change risk of defined benefit plans is greater than the market risk of 401(k) plans. To estimate the sensitivity of pension wealth to job change, they calculate defined benefit plan and 401(k) benefits for six benchmark careers, age 31 to 65, 31 to 42, 42 to 53, 53 to 65, 42 to 65, and 31 to 53. They calculate median retirement income and find that in all but one of these careers (age 53 to 65), median retirement income under 401(k)s exceeds the defined benefit plan median. For the full career workers (age 31 to 65) the authors also calculate the expected utility of the distribution of benefits under each plan and find that 401(k)s yield higher retirement incomes with less risk than defined benefit plans.

I will not comment further on this issue. Rather I will discuss some recent evidence on 401(k) alternative investment allocations and what is best for retirees. "Best" is measured by "risk-adjusted" wealth at retirement. (See Poterba, Rauh, Venti, and Wise, 2005 and 2006.) Wealth at retirement is not certain, but given the allocation of contributions, wealth depends on the rate of return, in particular the rate of return on stocks and bonds. The future realized returns on both are uncertain. Given assumptions about possible returns, a given allocation will yield a distribution of possible levels of wealth. The question we ask is how a person would value these distributions of wealth. We first obtain wealth distributions under different assumptions about portfolio allocation. Then for each portfolio allocation we calculate the expected "utility" of the distribution of wealth levels. The key idea is to give "appropriate" negative weight to possible bad wealth outcomes, which may be very unlikely to occur. We do this based on the commonly used constant relative risk aversion utility function. Then we ask what certain wealth level would be equivalent to the expected utility associated with a given distribution of wealth at retirement.

Table 1 shows these certainty equivalents for persons with two levels of risk tolerance--one is risk neutral and cares only about the expected value of the wealth distribution, the other is "moderately" risk averse. (7) Three portfolio allocation options are considered--100 percent investment in riskless index bonds (yielding 2.8 percent per year), 50 percent in bonds and 50 percent in stocks, and 100 percent in stocks. Certainty-equivalent wealth is shown for two groups: persons with a high school degree or less and persons with a college degree or more. Consider persons with a college degree or more who are risk neutral. In this case, the value of the distribution under any of the allocations is just the expected value of the possible wealth levels at retirement. Index bonds yield $265,000, 50 percent bonds and 50 percent stocks yields $509,000, and 100 percent stocks yields $1,042,000. The more interesting case is the valuation of the options for persons who are risk adverse, here "moderately" risk averse. In this case, the certainty equivalent of the 50 percent bond and the 50 percent stock allocation is $469,000, and the certainty equivalent of the 100 percent stock portfolio is $686,000. Thus, persons with this level of risk aversion would prefer the risky outcomes under the all-stock allocation than the certain outcome from riskless bonds, by a factor of over two and a half.

Further simulations show that retirees with very high risk aversion would prefer the 50 percent bond-50 percent stock portfolio, for example. But our results suggest that persons with modest risk aversion would be best off with investments over-weighted in stocks. Even if the return on stocks were 300 basis points less than past historical returns, the all stock portfolio seems to dominate the other options for persons with "moderate" risk aversion, based on our method of comparison. The conclusions for high school graduates are about the same as for college graduates. Our analysis also highlights the critical role of other sources of wealth, such as Social Security, defined benefit pension annuities, and saving outside retirement plans, in determining the expected utility cost of holding highly volatile assets in the retirement account. The greater other wealth, the less the effect on utility of uncertain 401(k) assets.

Life cycle funds are a recent innovation designed to relieve employees of the explicit choice of retirement account allocations. The key feature of these funds is that they reduce the proportion of assets in stocks and increase the proportion in bonds as employees approach retirement age. The funds were started in about 1994 and now hold about $50 billion in assets. Based on the typical allocation of stocks and bonds in such funds, we have compared the value of wealth distributions at retirement under these funds with wealth distributions under other allocations.

Results for selected allocations are shown in Table 2. (8) The method of comparison is exactly the same as described with respect to Table 1. Consider first the expected value of wealth at retirement, under the risk neutral heading. The expected value is $209,000 under a 100 percent government bond allocation, $818,000 under a 100 percent stock allocation, $451,000 under a life cycle stock and bond allocation, and $421,000 with a "fixed proportions" allocation. The fixed proportions allocation simply allocates contributions each year in fixed proportions to stocks and bonds. The proportions used here are selected to be the average proportions in a life cycle fund over a working life.

The more interesting results are, again, for retirees with moderate risk aversion. We find that for the "moderately" risk-averse retiree, the certainty equivalent of the 100 percent stock allocation is substantially greater than the certainty equivalent of the life cycle allocation--$584,000 compared to $403,000. The certainty equivalent of the fixed proportions allocation is about five percent less than the certainty equivalent of the life cycle allocation. We have not considered the administrative cost of managing a life cycle fund, and it is possible that this difference could be outweighed by such costs.

Again, the results depend on the degree of risk aversion that is assumed and on the rate of return assumed for stocks. But the results suggest that under the assumptions we find most plausible, an all-stock portfolio would be preferred to a life cycle allocation. On the other hand, it seems evident that life cycle funds may be a useful way for some participants in 401(k) plans to reduce risk.

Age, Earnings, and 401(k) Participation

The discussion above shows that 401(k) plans have expanded very rapidly since they were introduced in 1982. In addition, the projected growth in 401(k) plans is very large. The average 401(k) wealth of all persons who retired in 2000 is about $24,000 dollars, including retirees who are not covered by such a plan. By comparison, the average defined benefit wealth of persons 65 to 69 in 2000 is about $50,000. However, since the 401(k) program is relatively new, many people who retired in 2000 did not have long to contribute. Thus, it seems evident that wealth of future retirees under the personal account system will be much greater than the defined benefit system would have provided. Here I want to emphasize, however, what 1 see as the primary limitation of 401(k) plans (which has also been a limitation of defined benefit plans).

Low-income employees are much less likely than higher-income employees to be covered by 401(k) plans. Table 3 shows percent eligibility rates by annual earnings and by age in 2003, based on SIPP data. Eligibility rates do not differ much by age. But low-wage-earners are much less likely than employees with higher earnings to be covered by 401(k) plans. Over 80 percent of employees with incomes greater than $100,000 per year were eligible for a 401(k) plan; less than 50 percent of employees with earnings less than $25,000 per year were eligible.

Participation follows a similar pattern. Table 4 shows that over 70 percent of employees with annual earning over $100,000 participate; fewer than 30 percent of employees with earning less than $25,000 participate.


I believe that retirement saving under personal retirement accounts will be much greater than it has been under defined benefit plans. In addition, contrary to common belief, available evidence suggests that the market risk of personal accounts may pose less risk to employees than the job change risk of defined benefit plans. This is a question that I and my colleagues are now pursuing further. While a common perception also seems to be that stock investment poses the greatest risk to personal account participants, the work that my colleagues and I have done suggests that over a working life the typical employee would prefer the uncertain wealth accumulation ensuing from portfolio allocations heavily weighted to stocks. Even the distribution of wealth outcomes from life cycle funds, which give more weight to bonds as persons draw closer to retirement age, would likely be valued less than wealth outcomes from 100 percent stock allocations. Nonetheless, I believe that life cycle funds, and other products that would help employees to manage risk, should be encouraged. Such tools provide options that are likely to increase the advantages of personal accounts for some retirees. The latter conclusions are based on a rather stylized analysis and could be modified by a more thorough understanding of the risk tolerance of retirement savers. In addition, the analysis summarized above does not consider other non-market risks--like non-participation in a 401(k) plan, when eligible, or early withdrawal of plan assets at the time of job change--that may have substantial adverse affect of wealth accumulation at retirement. Finally, it seems to me that the greatest improvement to the personal account system would come from policies to hasten the penetration of 401(k) plans to low-income employees.


Johnson, Richard W., Leonard E. Burman, and Deborah I. Kobes. 2004. "Annuitized Wealth at Older Ages: Evidence from the Health and Retirement Study." Final Report to the Employee Benefits Security Administration. U.S. Department of Labor.

Kotlikoff, Laurence and David A. Wise. 1989a. "Pension Backloading, Wage Taxes, and Work Disincentives," in Laurence Summers (ed.), Tax Policy and the Economy, Volume 2. MIT Press.

______. 1989b. The Wage Carrot and the Pension Stick. W.E. Upjohn Institute.

______. 1989c. "Employee Retirement and a Firm's Pension Plan," in. Wise (ed.), The Economics of Aging. University of Chicago Press.

Poterba, James, Steven Venti, and David A. Wise. 2004. "The Transition to Personal Accounts and Increasing Retirement Wealth: Macro and Micro Evidence," in D.A. Wise, (ed.), Perspectives on the Economics of Aging. University of Chicago Press.

______. 2005. "Demographic Change, Retirement Saving, and Financial Market Returns." Working Paper Prepared for the 7th Annual Joint Conference of the Retirement Research Consortium, "Creating a Secure Retirement System," August 11-12, Washington, D.C.

Poterba, James M., Joshua Rauh, Steven F. Venti, and David A. Wise. 2005. "Utility Evaluation of Risk in Retirement Saving Accounts," In David A. Wise (ed.) Analyses in the Economics of Aging. University of Chicago Press.

______. 2006. "Life Cycle Asset Allocation Strategies and the Distribution of 401(k) Retirement Wealth." NBER Working Paper #11974. Also forthcoming in David A. Wise (ed.), Developments in the Economics of Aging, University of Chicago Press.

Samwick, Andrew and Jonathan Skinner. 2004. "How Will 401(k) Pension Plans Affect Retirement Income?" American Economic Review. 94, 1.

David A. Wise is the Stambaugh Professor of Political Economy at the John F. Kennedy School of Government at Harvard University. He is also affiliated with the National Bureau of Economic Research and the Hoover Institution at Stanford University. He has written extensively about the saving effect of personal retirement programs and more recently has been evaluating the implications of the rapid spread of these programs. He is currently engaged in analysis of the retirement incentives in public social security programs around the world. In addition, he has been analyzing the incentive effects and features of employer-provided health insurance programs, the financial implications of housing wealth for the elderly, social security reform, and other economics of aging issues.

This paper is a written version of a talk with the same title that I presented in Chicago at the annual meeting of the National Association of Business Economics on September 27, 2005. The paper is based largely on joint work with James Poterba at MIT, Josh Rauh at the University of Chicago, and Steven Venti at Dartmouth College.

(1) 401(k), 403(b) and 457(b) refer to sections of the Internal Revenue Code that provide for tax preferences for certain types of personal retirement accounts. Section 401(k) was the result of an amendment to the Code in 1978, but contributions to such plans did not begin until 1982.

(2) Detailed discussion of the job change risk under these plans can be found in Kotlikoff and Wise (1989a, 1989b, 1989c). Samwick and Skinner (2004) compare the risk under personal account and under defined benefit plans.

(3) A family's cohort assignment is based on the age of the male "head" of the household.

(4) Keogh plans are for self-employed people and are similar to IRAs.

(5) Details of the projection method are presented in Poterba, Venti, and Wise (2006).

(6) See Poterba, Venti, and Wise (2004). Contributions to defined benefit plans increased substantially after 2000 because of the fall in the stock market and the decline in long term interest rates, both of which reduced the value of plan assets relative to liabilities.

(7) With a CRRA risk aversion parameter of 2. For this analysis we focus on married couples. We assume that nine percent of a household's earnings are contributed to a 401(k) plan, including both employer and the employee contributions. We assume that the couple begins to participate in a 401(k) plan when the husband is 28 and that they contribute in every year in which the household has social security earnings until the husband is 63. Households do not make contributions when they are unemployed or when both members of the couple are retired or otherwise not in the labor force. When the husband is age 63, we assume that both members of the household retire if they have not already, and that contributions cease.

(8) The values for the 100 percent stock portfolio are different in Tables 1 and 2. Both tables pertain to two-person households, but the sample sizes are larger in Table 2 than in Table 1. In addition, there are some differences in specific procedures. The key message of the tables, however, is not affected by these differences.
RULES (in 1,000's of year 2000 dollars)

Portfolio Allocation HS or Less College or More

Risk Neutral
100% Riskless Index Bonds 199 265
50% Bonds, 50% Stocks 401 509
100% Stocks 852 1042

"Moderate" Risk Aversion
100% Riskless Index Bonds 199 265
50% Bonds, 50% Stocks 969 469
100% Stocks 523 686

Note: Assuming median value of SSW + DB Wealth
Source: Poterba, Rauh, Venti and Wise (2005).

RULES (in 1,000's of year 2000 dollars)

Allocation Rule HS or Less College or More

Risk Neutral
100% Government Bonds 163 209
100% Stocks 678 818
Life Cycle Stocks & Bonds 367 451
Fixed Proportions 340 421

"Moderate" Risk Aversion
100% Government Bonds 154 201
100% Stocks 442 584
Life Cycle Stocks & Bonds 315 403
Fixed Proportions 298 381

Source: Poterba, Rauh, Venti and Wise (2005).


Earnings Age
 <35 35-50 50-65 All

<$25K 47 46 47 47
25-50 67 69 67 68
50-100 75 77 79 77
>$100k 78 83 79 81
All 59 63 62 62

Source: Author's calculations for 2003 from SIPP data.


Earnings Age
 <35 35-50 50-65 All

<$25K 25 29 31 28
25-50 49 55 55 54
50-100 60 66 69 66
>$100k 62 74 73 72
All 40 48 49 46

Source: Author's calculations for 2003 from SIPP data.
COPYRIGHT 2006 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Wise, David A.
Publication:Business Economics
Geographic Code:1USA
Date:Apr 1, 2006
Previous Article:Enhancing Fed credibility.
Next Article:Improving incentives in health care spending: properly designed health spending accounts can be a major step.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters