Retirement Security in the United States: Current Sources, Future Prospects, and Likely Outcomes of Current Trends

Retirement Security in the United States Current Sources, Future Prospects, and Likely Outcomes of Current Trends

Jack VanDerhei Craig Copeland Dallas Salisbury A Publication of the Employee Benefit Research Institute

Copyright © 2006 by the Employee Benefit Research Institute–Education and Research Fund (EBRI-ERF). All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing.

Library of Congress Cataloging-in-Publication Data VanDerhei, Jack L. Retirement security in the United States : current sources, future prospects, and likely outcomes of current trends / Jack VanDerhei, Craig Copeland, Dallas Salisbury. p. cm. Includes bibliographical references. ISBN 0-86643-100-4 (alk. paper) 1. Retirement income--United States. 2. Social security--United States. I. Copeland, Craig, 1968- II. Salisbury, Dallas L. III. Title. HD7125.V36 2006 331.25’20973--dc22 2006024780

Jack VanDerhei is a faculty member at Temple University’s School of Business and Management (Department of Risk, Insurance, and Healthcare Management) and the Research Director of the EBRI Fellows Program. He is the editor of Benefits Quarterly and author of numerous articles on retirement plans and policy. Craig Copeland is a senior research associate with the Employee Benefit Research Institute and director of the EBRI’s Social Security Reform Evaluation Research Program. Dallas Salisbury is president and CEO of the nonpartisan Employee Benefit Research Institute (EBRI), which was founded in Washington, DC, in 1978 and provides objective information regarding the employee benefit system and related financial security issues. Mr. Salisbury joined EBRI at its founding in 1978, and prior to joining EBRI held fulltime positions with the Washington State Legislature, the U.S. Department of Justice, the Employee Benefits Security Administration of the U.S. Department of Labor (formerly known as the PWBA), and the Pension Benefit Guaranty Corporation (PBGC).

The Employee Benefit Research Institute (EBRI) was founded in 1978. EBRI’s Education and Research Fund (ERF), which performs the Institute’s charitable, educational, and scientific functions, is a tax-exempt 501(c)(3) organization supported by contributions and grants (it is not a foundation). EBRI’s mission is to contribute to, to encourage, and to enhance the development of sound employee benefit programs and sound public policy through objective research and education. EBRI is the only private, nonprofit, nonpartisan, Washington, DC-based organization committed exclusively to public policy research and education on economic security and employee benefit issues. EBRI’s membership includes a cross-section of pension funds; businesses; trade associations; labor unions; health care providers and insurers; government organizations; and service firms. Its Web site is www.ebri.org.

Executive Summary



Social Security (OASDI) is the most prevalent source of income for those in the traditionally defined “retirement ages” of 65 or older. Approximately 90 percent of Americans 65 or older had Social Security as a source of income in 2004.



When examining those elderly with incomes in the top half in 2004, pension income is received by 63 percent in the top quartile and 51 percent in the third quartile. In contrast, less than 20 percent of those in the bottom two income quartiles received a pension in 2004.



For those ages 65–69 in the top quartile of income, 54 percent received pension income, compared with 72 percent of those ages 85 or older. This result is also found in the third income quartile, as 42 percent of those ages 65–69 had pension income, compared with 52 percent of those ages 85 or older. In the lowest two income quartiles, the differences are much smaller, but those 85 and older still have a higher (albeit very slight) incidence of pension income relative to those ages 65–69.



The percentage of the elderly’s income coming from Social Security, public pensions, private pensions, and other sources has remained virtually constant from 1987 to 2004, at approximately 40 percent (Social Security), 9 percent (public pensions), 10 percent (private pensions), and 4 percent (other sources), respectively.



When the elderly population is broken down into income quartiles, pensions account for 5 percent or less of the total income for those among the bottom two quartiles in 2004. Social Security is the overwhelmingly dominant source of income for individuals in the bottom two quartiles, accounting for 92 percent of income for those in the bottom quartile and 84 percent for those in the second quartile. In the upper two quartiles, pensions account for a much larger percentage Executive Summary

III

IV

Retirement Security in the United States

than they did in the bottom two quartiles, but private pensions still accounted for only 12 percent of the income for those in both of the upper two quartiles and public pensions accounted for 7 percent for those in the third quartile and 13 percent for those in the fourth (highest) quartile. 

On average, the elderly had what many would consider rather modest incomes in 2004: the mean (average) annual income was $22,918 and the median (midpoint, half above and half below) annual income was $14,650. Data show that the amount of income the elderly received decreased as their age increased. Those ages 65–69 had a mean income of $28,314 in 2004, compared with $17,385 for those 85 or older. The average income for the third income quartile was $19,487; only those in the highest (fourth) income quartile had what many would consider sizable income: $55,790.



The mean pension amount among those receiving a pension was $10,354, and the median was $6,636. The mean and median public-pension amounts were significantly higher, at $19,978 and $15,600, respectively.



Assuming that each worker contributes an additional 5 percent of compensation from 2003 until retirement age to supplement his or her Social Security and tax-qualified retirement plans, approximately 30 percent of the simulated life-paths for the lowest income quartile for those in the 1936-1940 birth cohort would be expected to have sufficient retirement resources. However, at least 85 percent of the simulated lifepaths for the third or fourth income quartiles for those in the 1961-1965 birth cohort would be sufficient. For each birth cohort, the lower-income quartiles have more risk of insufficient retirement income than their higher-paid counterparts. Moreover, single females tend to exhibit more vulnerability than single males, while families are typically the least vulnerable.



In 1987, 28.2 percent of the American workforce was age 25–34. By 2004, this had declined to 21.3 percent. In contrast, 15.0 percent of workers were age 45–54 in 1987, which grew to 22.7 percent by 2004. Furthermore, 20.4 percent of workers in 1987 were under age 25, compared with just over 15 percent in 2004, while the percentage of workers age 55 or

Employee Benefit Research Institute

older increased from 13.5 percent in 1987 to 17.1 percent in 2004. 

In 1965, there were 4.0 Social Security-covered workers for every Social Security beneficiary. In 2005, this number had fallen to 3.3 workers per beneficiary. Once all of those in the baby boom generation reach age 65 in about 2030, there will be just over two workers for every beneficiary. This ratio is then projected to fall below two by 2065.



The percentage of American workers participating in an employment-based retirement plan (either a defined benefit or defined contribution plan) has been fairly constant, increasing slightly from 1987 to 2004. Among all workers, the percentage who participated in a plan increased from 37.6 percent in 1987 to 41.9 percent in 2004, with even higher levels from 1998 to 2001. When focusing more specifically on private-sector wage and salary workers (excluding self-employed workers) ages 21–64, the participation rate increased from 39.8 percent in 1987 to 43.0 percent in 2004. Retirement plan rates of participation of full-time, full-year workers and public-sector workers were higher but unchanged from 1987 to 2004.



While the percentage of workers who participate in an employment-based retirement plan has remained relatively constant, the type of plan that private-sector workers are participating in has changed dramatically. In 1979, 62 percent of private-sector workers who were participating in an employment-based retirement plan had a defined benefit (pension) plan solely, with another 22 percent having a defined benefit plan along with a defined contribution (401(k)-type) plan. Only 16 percent of the participants had a defined contribution plan solely. However, by 2004, the percentages with only one plan type had switched, as 63 percent of the participants had a defined contribution plan solely and 10 percent had a defined benefit plan solely. Those participants with both types increased slightly, to 27 percent.



Furthermore, the method for determining benefits within the so-called “traditional” pension (defined benefit) plans has also been changing during this transformation toward defined contribution plans. Most defined benefit plans have tradi-

V

VI

Retirement Security in the United States

tionally used a career average pay or final-average pay method to determine benefits. However, other arrangements have become more common among defined benefit plans, as 5 percent of defined benefit plans (and more significantly, 25 percent of participants) were in so-called “hybrid” defined benefit plans—most commonly a cash balance plan—in 2003. 

Along with the change in method of benefit determination, the allowable payout methods also changed. The percentage of participants in a defined benefit that had a lump-sum option increased from 15 percent in 1995 to 48 percent in 2002. The net result has been a significant reduction in the number of remaining defined benefit pension plans that pay out a benefit in the form of a lifetime annuity.



In 2001, 75 percent of defined contribution-eligible participants participated in an offered plan. This was higher than the 74 percent level in 1995, but below the peak of 77 percent in 1998. The participation rate is strongly affected by the income of the individual, with the likelihood of participation increasing with income. Specifically, in 2001, 53 percent of eligible workers with family incomes below $10,000 participated in the offered defined contribution plan, compared with 85 percent of those with family incomes of $100,000 or more.



In 2003, the average contribution to salary reduction plans (i.e., 401(k), 403(b), or thrift savings plans) was 7.5 percent of annual earnings. This is an increase from 6.6 percent to such plans in 1988.



In 2004, 46 percent of total 401(k) plan assets were invested in equities, 10 percent in balanced funds, 15 percent in company stock, 10 percent in bond funds, 4 percent in money funds, and 12 percent in guaranteed investment contracts and stable value funds. These numbers have essentially been at these levels since 1996. Furthermore, as the age of the participant increases the percent in equities declines (50.9 percent for those in their 40s compared with 36.5 percent for those in their 60s).



The percentage of individuals who took a lump-sum distribution from an employment-based retirement plan and rolled

Employee Benefit Research Institute

over the entire amount of the distribution to a tax-qualified savings vehicle (i.e, individual retirement accounts, another job’s plan) has increased with time before leveling off. Among those distributions taken before 1980, 21 percent were entirely rolled over. This percentage increased to 48 percent for those distributions taken from 1994–1998, before falling to 45 percent for distributions taken from 1999–2003. However, the percentage of individuals who have taken a lumpsum distribution and used any proportion for consumption has continued to decrease, from 36.4 percent of distributions taken before 1980 to 23.4 percent for those taken from 1999–2003. 

The amount of the distribution affects the likelihood of an individual rolling over the distribution: The larger the amount, the more likely it will be rolled over. For distributions of $1–$499 (2003 dollars), 19.9 percent of lump-sum recipients rolled over their entire distribution. This increased with the amount of distribution to 68.1 percent for distributions of $50,000 or more.



In 1990, the projected 75-year actuarial balance for the OldAge, Survivors, and Disability Insurance (OASDI) program was –0.91 percent of taxable OASDI payroll, meaning that the OASDI or Social Security tax would need to be increased by 0.91 percentage points to achieve actuarial balance over the next 75 years from 1990. This 75-year deficit grew to more than 2 percent of taxable payroll before improving to just under 2 percent by 2005. However, this would still be a significant increase of the current taxes—or a corresponding cut in benefits—in order to cover this deficit.



On top of the projected deficit, the amount of preretirement income a worker can expect at retirement from Social Security at specific ages of commencing benefits is going to decline due to the scheduled increases in the normal retirement age. For those age 65 in 2005 and starting to receiving benefits in that year and having made the equivalent of the average wage index over the course of their working career, the replacement rate of income just prior to turning age 65 from Social Security is 42.2 percent. For those turning age 65 in 2025 with the same earnings, the expected replacement rate is 36.3 per-

VII

VIII

Retirement Security in the United States

cent. The replacement rates from Social Security are lower for higher earners and higher for lower earners. 

The Medicare Hospital Insurance Trust Fund is estimated to be facing an actuarial balance of –3.09 percent of taxable payroll, which is a significant increase from the projected balance in 2002 of –2.02 percent. This means that the payroll tax for Medicare would need to be increased by 3.09 percentage points—or benefits would have to be cut by an equivalent amount—in order for the revenues to match expenses for the program over the next 75 years. This level of tax increase would more than double the current payroll tax rate for Medicare.



The projected growth in these expenditures could present serious issues for the federal and state governments. According to a Congressional Budget Office (CBO) study, depending on the spending path of expenditures for Medicaid and Medicare, these expenditures could amount to from 5.3 percent of gross domestic product (GDP) to 21.9 percent of GDP by 2050. For comparison, Social Security expenditures are projected to account for between 6.3 percent to 6.6 percent of GDP by 2050 under the same sending path assumptions. Consequently, given the funding challenges facing government programs and employers, individuals are also likely to be faced with increased challenges of funding a consistent level of expenditures in retirement. Therefore, significantly more preparation and thought must occur at each level in order to maintain a standard of living of all retired Americans.



The vast majority of cohorts would need to increase their savings rate between 10 percent and 20 percent of the rate needed if all defined benefit participants (other than those in cash balance plans) were assumed to receive an annuity.



Choice complexity occurs when taxpayers are given numerous subsidies but may only use one or a few of them. This type of complexity has grown into the large number of savings tools for health care and retirement described above. Although more choices may be advantageous to some individuals for tax planning and/or maximum contributions, this increased complexity will certainly have attendant costs with respect to learning about the various requirements for each.

Employee Benefit Research Institute



An EBRI/ICI simulation conducted to assess the potential value to employees of 401(k) plan sponsors adopting automatic enrollment found significant increases in average account balances for low-income employees regardless of the choice of the default contribution rate and asset allocation. As the default contribution rate is increased from 3 to 6 percent of compensation and as the default asset allocation is changed from a very conservative selection to one that contains an ageappropriate mix of equities and bonds, the average account balances for all income categories increase.



The median annual contribution rate needed to financially indemnify a participant in a career-average defined benefit pension plan whose plan was frozen in 2006 is simulated to be 11.6 percent, assuming a 4 percent rate of return. An contribution rate of 18.8 percent would be sufficient to cover 75 percent of the employees covered by this type of plan. The median rate for a final-average plan is larger, as expected: 13.5 percent, and the threshold rate for the 75th percentile increases to 21.0 percent. Cash balance plans have a median contribution rate of 4.6 percent, with a 75th percentile threshold rate of 6.3 percent using the current interest credits. These values increase to 5.7 percent and 7.3 percent if, instead, the cash balance plans are assumed to credit interest at the intermediate long-term assumption for the interest rate of the Treasury special public-debt obligation bonds issuable to the OASDI trust funds, as specified in the 2005 Trustees of the OASDI Trust Funds Report (5.8 percent). If the rate of return assumption is increased to 8 percent nominal, the median contribution rate for a career-average defined benefit plan is 6.6 percent. A contribution rate of 14.8 percent would be sufficient to cover 75 percent of the employees covered by this type of plan. The median for a final-average plan is 8.1 percent and the 75th percentile threshold increases to 16.0 percent. Cash balance plans have a median contribution rate of 2.7 percent, with a 75th percentile threshold of 4.5 percent using the current interest credits. These values increase to 3.1 percent and 5.2 percent if the cash balance plans are assumed to credit interest at 5.8 percent.

IX

TABLE OF CONTENTS Executive Summary ........................................................................... I Figures ........................................................................................... 1 I. II.

Introduction ......................................................................... 7 Current Retirement Income Sources...................................... 11 Incidence .......................................................................................13 Relative Amounts............................................................................16 Current Status of Retirees ................................................................30

III.

Changes and Factors Affecting Future Retirement Income Prospects ........................................................................... 41 Aging of United States Population .....................................................43 Retiree Health Care .........................................................................44 Retirement Plan Participation ...........................................................44 Individual Choices Affecting Retirement Income..................................50

IV.

Cash Balance Plans ............................................................ 63 Fundamental Economic Distinction Between Final-Average and Cash Balance Plans ...............................................................................67 Potential Advantages: Cash Balance vs. Final-Average Plans ................68 Potential Advantages: Cash Balance vs. Defined Contribution Plans ......71 Potential Limitations of a Conversion From a Defined Benefit to a Cash Balance Plan .................................................................................72 Key Issues ....................................................................................73 Pension Protection Act of 2006 ........................................................75

V.

Big Picture Overview of How Retirement “Security” Will be Delivered in the Future ........................................................ 77 Potential Modifications to Social Security, Medicare and Medicaid ........79 The Impact of Moving to a Reduced Social Security Benefits Scenario ...86 The Impact of Assuming Lump-Sum Distributions Are Offered to All Defined Benefit Plan Participants at Retirement and They Are Always Chosen ..........................................................................................86

VI.

Are Employees Overwhelmed With Choices? .......................... 91 A Wide Variety of Savings Tools for Health Care and Retirement ...........93 Options for Consolidation and Simplification ......................................99

VII.

An Example of How Employers Can Improve Retirement Security for Their Employees .............................................. 101 Legislative/Regulatory Environment for Auto-Enrollment ....................103 Cost/Benefit Analysis for Employers Considering Auto-Enrollment .......104 Simulation Results ........................................................................105

VIII.

Defined Benefit Plan Freezes ............................................. 109 Financial Consequences of a Pension Freeze ....................................111 Analyzing the Financial Consequences of a Pension Freeze ................114 Overall Results by Plan Type ..........................................................115 Results by Age and Plan Type ........................................................116 Results by Tenure With the Current Employer at the Time of the Pension Freeze and by Plan Type ....................................................116 Results by Remaining Years to Retirement After Leaving Job With the Current Employer and by Plan Type ...........................................121 Results by Tenure With Employer After the Pension Freeze Until Job Change and by Plan Type.........................................................126

IX.

Conclusion ....................................................................... 131

X.

References ....................................................................... 135

Appendix ..................................................................................... 141 Endnotes ..................................................................................... 149

Figures

1

Page 14

Figure 1, Percentage of Elderly With Various Sources of Income, Various Years, 1987–2004

Page 15

Figure 2, Percentage of the Elderly With Various Sources of Income, by Age, 1987

Page 17

Figure 3, Percentage of the Elderly With Various Sources of Income, by Age, 2004

Page 18

Figure 4, Percentage of Elderly With Various Sources of Income, by Income Quartile, 2004

Page 19

Figure 5, Percentage of 65–69 Year Olds’ Income With Various Sources of Income, by Income Quartile, 2004

Page 20

Figure 6, Percentage of 85-or-Older-Year-Olds With Various Sources of Income, by Income Quartile, 2004

Page 21

Figure 7, Percentage of the Elderly’s Income From Major Sources, Various Years 1987–2004

Page 23

Figure 8, Percentage of the Elderly’s Income From Major Sources, by Age, 1987

Page 24

Figure 9, Percentage of the Elderly’s Income From Major Sources, by Age, 2004

Page 25

Figure 10, Percentage of the Elderly’s Income From Various Sources, by Income Quartile, 2004

Page 26

Figure 11, Percentage of Americans’ Income From Various Sources, Ages 64–69, by Income Quartile, 2004

Page 27

Figure 12, Percentage of Americans’ Income From Various Sources, Ages 85 or Older, by Income Quartile, 2004

Page 29

Figure 13, Elderly Americans’ Mean and Median Annual Income, Mean and Median Annual Pension Income for those Having Such Pensions, by Age and Income Quartile, 1987, 1994, and 2004

Page 36

Figure 14, Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 75% Chance of Covering Basic Retirement Expenses

Page 37

Figure 15, Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 90% Chance of Covering Basic Retirement Expenses

Figures

3

4

Retirement Security in the United States

Page 38

Figure 16, Percentage of Retirees Estimated to Have Sufficient Retirement Income/Wealth by Saving 5% of Compensation Each Year From 2003 Until Retirement

Page 43

Figure 17, Age Distribution of the American Workforce, 1987– 2004

Page 45

Figure 18, Ratio of Old-Age, Survivors, and Disability Insurance (OASDI)-Covered Workers Per OASDI Beneficiary, 1965–2075

Page 46

Figure 19, Average Increase in Total Retiree Health Costs for Private-Sector Firms With 1,000 or More Employees, 2002–2004

Page 47

Figure 20, Percentage of All Large Firms (200 or More Workers) That Offer Health Benefits Also Offering Retiree Health Benefits, 1988–2005

Page 48

Figure 21, Percentage of Various Work Forces Who Participate in an Employment-Based Retirement Plan, 1987–2004

Page 49

Figure 22, Distribution of Private-Sector Workers Who Are Participating in an Employment-Based Retirement Plan, by Plan Type, 1979–2004

Page 51

Figure 23, Percentage of Private-Sector Defined Benefit Plans Who Are Hybrid Plans and Participants Who Are in Hybrid Plans, 2001– 2003

Page 52

Figure 24, Percentage of Private-Sector Defined Benefit Plan Participants Having a Lump-Sum Distribution Option Available, 1995, 1997, and 2002

Page 53

Figure 25, Participation Rates of Family Heads Eligible for an Employment-Based Defined Contribution Plan, 1995, 1998, and 2001

Page 55

Figure 26, Average Percentage of Annual Earnings Contributed to a Salary Reduction Plan, Civilian Nonagricultural Wage and Salary Workers, Age 16 and Over, 1988, 1993, 1998, and 2003

Page 56

Figure 27, Proportion of Lump-Sum Recipients Using Entire Portion of Their Most Recent Distribution for Tax-Qualified Financial Savings, by Year of Receipt, Civilians Age 21 and Over

Page 57

Figure 28, Proportion of Lump-Sum Recipients Reporting Using Any Portion of Their Most Recent Distribution for Consumption, by Year of Receipt, Civilians Age 21 and Over

Page 58

Figure 29, Proportion of Lump-Sum Recipients Using Entire Portion of Their Most Recent Distribution for Tax-Qualified Financial Savings, by the Amount of the Most Recent Distribution, Civilians Age 21 and Over

Page 60

Figure 30, Percentage of Former Participants in an EmploymentBased Retirement Plan Who Either Took a Lump-Sum Distribution or Retained Their Benefits in Their Previous Job’s Plan, 2003

Employee Benefit Research Institute

Page 61

Figure 31, Percentage of Individuals Who Had a Retirement Plan at a Previous Job Who Took a Lump-Sum Distribution vs. Leaving the Assets in the Plan, by the Value of the Distribution and the Current Plan Balance, 2003

Page 66

Figure 32, Illustration of a Conversion From a Hypothetical Traditional Final-Average Defined Benefit Plan to a Hypothetical ServiceWeighted Cash Balance Plan (Without Transition Credits) at Age 55

Page 70

Figure 33, Hypothetical Percentage Increases in Annual Benefits at Normal Retirement Age. Cash Balance vs. Final-Average Plan: Impact of Job Tenure

Page 80

Figure 34, Actuarial Balance of the Federal Old-Age and Survivors Insurance and Disability Insurance (OASDI) Trust Funds, 1990– 2005

Page 81

Figure 35, Replacement Rate of the Social Security Retirement Benefit for Those Retiring at Age 65 From 2005–2025, by Various Steady Earnings Histories

Page 83

Figure 36, Actual and Projected Percentage Increases in Medicare and Medicaid Personal Health Care Expenditures, 1998–2014

Page 84

Figure 37, Medicare Hospital Insurance (HI) Trust Fund Actuarial Balance, 2002–2005

Page 85

Figure 38, Projected Percentage of Gross Domestic Product (GDP) of Medicare and Medicaid and Social Security Expenditures, by Spending Path Assumption, 2010, 2030, and 2050

Page 87

Figure 39, Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 75% Chance of Covering Basic Retirement Expenses

Page 89

Figure 40, Increase in Median Percentage of Additional Compensation That Must Be Saved Annually Until Retirement for a 75% Chance of Covering Simulated Expenses, as a Result of Assuming All Defined Benefit Participants Take LSDs at Retirement

Page 117

Figure 41, Cumulative Distribution Function of the Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type (assumes a 4% annual rate of return)

Page 118

Figure 42, Cumulative Distribution Function of the Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type (assumes 8% annual rate of return)

Page 119

Figure 43, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Current Age (assumes 4% annual rate of return)

5

6

Retirement Security in the United States

Page 120

Figure 44, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Current Age (assumes 8% annual rate of return)

Page 122

Figure 45, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Years Already Worked on the Job With the Current Employer (assumes 4% annual rate of return)

Page 123

Figure 46, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Years Already Worked on the Job With the Current Employer (assumes 8% annual rate of return)

Page 124

Figure 47, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years to Retirement after Leaving Job With the Current Employer (assumes 4% annual rate of return)

Page 125

Figure 48, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years to Retirement after Leaving Job With the Current Employer (assumes 8% annual rate of return)

Page 128

Figure 49, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type, and Remaining Years on the Job With Current Employer (assumes 8% annual rate of return)

Page 129

Figure 50, Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type, and Remaining Years on the Job With Current Employer (assumes 4% annual rate of return)

I

Introduction

7

A

s the nation begins the demographic shift of boomers moving towards retirement the population over age 65 will climb from just over 13 percent to nearly 22 percent. This shift will bring with it a progressive increase in attention to issues related to delaying retirement, retiree health availability and affordability, long term care, and how to produce income in retirement. There will be a tendency to rely upon generalizations or hypotheticals as opposed to detailed analysis. This report is intended to show the inherent risk in not undertaking detailed analysis for concerned parties. In Chapter 2 we review the current retirement income sources and provide detailed analysis of both the incidence and relative amounts. In addition to looking at merely the income levels currently generated, we summarize our previous simulation work projecting both retirement income and basic retirement expenses and potential uncovered health care costs in retirement. The primary objective of this analysis is to combine the simulated retirement income and wealth with the simulated retiree expenditures to determine how much each family unit would need to save today (as percentage of their current wages) to maintain a prespecified “comfort level” (i.e., confidence level) that they will be able to able to afford the simulated expenses for the remainder of the lifetime of the family unit (i.e., death of second spouse in a family). We find that if the status quo is not modified, the results are extremely depressing for those close to retirement age, especially single women. In Chapter 3 we review the changes and factors affecting future retirement income prospects. While the aging of the US population is a commonly recognized problem for the future of Social Security, we provide additional details on this topic as well as the attendant problems of the increasing expense (and decreasing private employer coverage) of retiree health care. Given the increasing trend away from defined benefit towards defined contribution retirement plan coverage among private employers, we pay particular attention not only to the retirement plan participation trends, but also the individual choices affecting retirement income. In addition to the participation decision, this includes employee choice with respect to contribution rates, asset allocation and preservation of balances at job change.

Introduction

9

10

Retirement Security in the United States

Chapter 4 reviews the recent developments of the retirement plan design that many feel may be the last chance for continued survival of defined benefit plans among private plan sponsors: cash balance plans. With Congress on the verge of deciding which form of cash balance design (if any) will receive special treatment for prospective conversions, we provide a detailed discussion of the fundamental economic distinction between final average and cash balance plans as well as a cost/benefit analysis of cash balance plans versus both final average defined benefit plans and defined contribution plans. We conclude this section with a discussion of the potential limitations of a conversion from a defined benefit plan to a cash balance plan and a review of the key issues at the time of this writing. Chapter 5 provides an overview of how retirement security (or an approximation thereof ) will be delivered in the future. This material reviews the potential modifications to Social Security, Medicare and Medicaid and their impact on future retirees. It also summarizes the results of earlier work on the impact of employees voluntarily choosing lump sum distributions from their defined benefit plans. This is followed by the topic of whether employees are being overwhelmed with choices in Chapter 6. As the evolution from defined benefit to defined contribution plans (especially those of the salary deferral or 401(k) type) has continued over the last two decades, many policy makers have been concerned that employees may not be taking advantage of the employer sponsored retirement plans soon enough (or in some cases at all) to end up with an adequate retirement income. A relative new procedure known as automatic enrollment or negative election has been adopted by some 401(k) sponsors in an attempt to use inertia to the employee’s advantage by initially enrolling them in the 401(k) plan and then giving them the option to opt out at their discretion. The recent trends with respect to automatic enrollment is reviewed in Chapter 7 and recent simulation work conducted by EBRI and ICI shows that this may be the singular bright spot with respect to future plan design trends for improving retirement security, especially among the low income workers. The dawn of the new year in 2006 began with an increase of press coverage of a decision among plan sponsors that had actually been quite prevalent in recent years: freezing defined benefit plan accruals for new and/or existing employees. Chapter 8 provides a detailed analysis of how such activity is likely to impact existing employees as a function of plan type and employee demographics.

II

Current Retirement Income Sources Incidence and Relative Amounts

11

Incidence1

S

ocial Security (OASDI) is the most prevalent source of income for those in the traditionally defined “retirement ages” of 65 or older. Approximately 90 percent of Americans 65 or older had Social Security as a source of income in 2004 (Figure 1). This percentage has been around 90 percent since at least 1987. However, there has been a slight decline in this trend from 91 percent in 1987 to 88 percent in 2004. The percentage having investment income has decreased from 70 percent in 1987 to 56 percent in 2004. Furthermore, those receiving income from other sources have trended downwards from 16 percent in 1987 to 10 percent in 2004. The only source with an overall upward trend was earnings, as 18 percent of the elderly had wages and salary income in 2004 compared with 15 percent in 1987.

The percentage of those 65 or older who received a pension payment from a public-sector sponsor remained constant at 11 percent from 1987 to 2004. The percentage of these same individuals who received a pension payment from a private-sector company or union sponsor increased from 21 percent in 1987 to 25 percent in 1999 before falling to 23 percent in 2002 and remaining at that level in 2004. Consequently, for the last approximately 20 years, only a third of those 65 or older received pension payments from a public- or private-sector sponsor. This even includes regular payments from individual retirement accounts (IRAs), Keoghs, and 401(k) accounts. However, it does not include prior or other lump-sum withdrawals from these types of plans which would fall under investment income, whose incidence has been falling. Therefore, the vast majority of elderly Americans have not been dependent upon a pension from a public- or private-sector sponsor; instead, they are primarily dependant upon a pension from Social Security. Examining income sources by the age of the individual, in 1987 the “younger elderly” were the ones more likely to be receiving a pension (Figure 2). Among those ages 65–69, 35 percent received a pension (public or private sponsor) compared with 22 percent of the “older elderly” (those 85 or older). However, in 2004, the oldest elderly were Current Retirement Income Sources

13

11%

21%

OASDI

1987

15%

91%

70% 68%

14%

Year

63%

12%

11%

23%

56%

11%

88%

Investment Income

2002

16%

89%

Public Pension

11%

25%

1999

17%

89%

Private Pension

11%

24%

1994

15%

Earnings

16%

91%

Source: Employee Benefit Research Institute estimates from the 1988, 1995, 2000, 2003, and 2005 March Current Population Survey.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

56%

11%

Other

2004

18%

23%

Figure 1 Percentage of Elderly With Various Sources of Income, Various Years, 1987–2004

10%

14 Retirement Security in the United States

OASDI

All

16%

70%

11%

21%

15%

91%

Earnings

65-69

75-79

14%

69%

10%

21%

9%

95%

Public Pension

Age

14%

72%

12%

70-74

15%

23%

Private Pension

16%

71%

12%

26%23%

86%

93%

Source: Employee Benefit Research Institute estimates from the 1988 March Current Population Survey.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Investment Income

80-84

17%

67%

11%

18%

5%

95%

Figure 2 Percentage of the Elderly With Various Sources of Income, by Age, 1987

Other

20%

63%

85 or older

8%

14%

3%

90%

Employee Benefit Research Institute 15

16

Retirement Security in the United States

slightly more likely to be receiving a pension, with 33 percent of those age 85 or older receiving a pension, compared with 31 percent of those ages 65–69 (Figure 3). Furthermore, 37 percent of those ages 80–84 received a pension in 2004. Having more sources of income typically would lead to higher levels of income. Therefore, by grouping the elderly into income quartiles, the higher-income individuals can be separated from the lower-income individuals. When examining those elderly with incomes in the top half in 2004, pension income clearly becomes a more prevalent source of income, and is received by 63 percent in the top quartile and 51 percent in the third quartile (Figure 4). In contrast, less than 20 percent of those in the bottom two income quartiles received a pension in 2004. The greater likelihood of the older elderly having received pension income in 2004 becomes even more pronounced when examined by income quartiles. For those ages 65–69 in the top quartile of income, 54 percent received pension income, compared with 72 percent of those ages 85 or older (Figures 5 and 6). This result is also found in the third income quartile, as 42 percent of those ages 65–69 had pension income, compared with 52 percent of those ages 85 or older. In the lowest two income quartiles, the differences are much smaller, but those 85 and older still have a higher (albeit very slight) incidence of pension income relative to those ages 65–69.

Relative Amounts The section above documented the percentage of elderly with various sources of income, but it did not show the percentage of the elderly’s income that these sources represent. This section describes these relative amounts of income coming from the major sources described above. The percentage of the elderly’s income coming from Social Security, public pensions, private pensions, and other sources has remained virtually constant from 1987 to 2004, at approximately 40 percent (Social Security), 9 percent (public pensions), 10 percent (private pensions), and 4 percent (other sources), respectively (Figure 7). The percentage of income that earnings represented increased from 15 percent in 1987 to 23 percent in 2004, while the percentage of income derived from investment income decreased from 24 percent to 12 percent. These results imply that public-sector pensions are much more generous than privatesector pensions, as twice as many of the elderly receive a private pension

OASDI

All

10%

56%

11%

23%

18%

88%

Earnings

65-69

10%

57%

11%

20%

32%

83%

Private Pension

70-74

Age

10%

57%

11%

24% 20%

89%

55%

Public Pension

75-79

11%

24%

12%

90%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

9%

Investment Income

80-84

12%

56%

11%

26%

6%

91%

Figure 3 Percentage of the Elderly With Various Sources of Income, by Age, 2004

Other

12%

52%

85 or older

9%

24%

4%

92%

Employee Benefit Research Institute 17

All

11%

23%

OASDI

18%

88%

56%

Earnings

10% 3% 1%

33%

13% 6% 4%

14%

47%

7%

Public Pension

25%-50% Income Quartile

Private Pension

Lowest 25%

3%

76%

97%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

13%

38%

65%

9%

26%

37%

43%

79%

Other

Highest 25%

86%

Investment Income

50%-75%

18%

94%

Figure 4 Percentage of Elderly With Various Sources of Income, by Income Quartile, 2004

12%

18 Retirement Security in the United States

83%

OASDI

All

11%

20%

32%

57%

Earnings

10%

5% 3% 1%

31%

13%

Private Pension

Lowest 25%

73%

93%

7%

Public Pension

Year

25%-50%

4%

12%

15%

50%

89%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

11%

31%

8%

23%

31%

59%

77%

Other

Highest 25%

81%

Investment Income

50%-75%

33%

62%

Figure 5 Percentage of 65-69 Year Olds' Income With Various Sources of Income, by Income Quartile, 2004

11%

Employee Benefit Research Institute 19

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

All

24%

9%

52%

Earnings

12% 3% 2%

27% 19%

3%

25%-50%

1%

15%

43%

8%

Public Pension

Income Quartile Private Pension

Lowest 25%

1%

80%

97%

65%

12%

40%

8%

85%

Other

Highest 25%

32%

40%

12%

89%

Investment Income

50%-75%

4%

97%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

OASDI

4%

92%

Figure 6 Percentage of 85-or-Older-Year-Olds With Various Sources of Income, by Income Quartile, 2004

15%

20 Retirement Security in the United States

Earnings

10%

15%

15%

19%

22%

23%

20%

OASDI

30%

40%

Private Pension

39%

44%

39%

42%

41%

60%

Public Pension

50%

8%

10%

10%

9%

8%

80%

9%

9%

Investment Income

70%

9%

10%

10%

Source: Employee Benefit Research Institute estimates from the 1988, 1995, 2000, 2003, and 2005 March Current Population Survey.

0%

1987

1994

1999

2002

2004

24%

14%

12%

Other

90%

17%

19%

Figure 7 Percentage of the Elderly's Income From Major Sources, Various Years 1987–2004

100%

5%

5%

4%

3%

3%

Employee Benefit Research Institute 21

22

Retirement Security in the United States

relative to those receiving a public pension, but the total private pension income of the elderly was only slightly above the income from public pensions (10 percent compared with 9 percent). In 1987, the younger elderly had a higher percentage of income coming from pensions than the older elderly. Those ages 65–69 received 10 percent and 9 percent of their income from private and public pensions, respectively, compared with 5 percent and 6 percent for those ages 85 or older (Figure 8). In contrast, in 2004, those 65–69 received the lowest percentage of their income from pensions, at 9 percent for each source (Figure 9). Those ages 80–84 received the highest percentage of income from pensions, with 12 percent from private pensions and 11 percent from public pensions. In both years, the percentage of income coming from Social Security and investment income increased with age, as only the percentage of income coming from earnings decreased. When the elderly population is broken down into income quartiles, pensions account for 5 percent or less of the total income for those among the bottom two quartiles in 2004 (Figure 10). Social Security is the overwhelmingly dominant source of income for individuals in the bottom two quartiles, accounting for 92 percent of income for those in the bottom quartile and 84 percent for those in the second quartile. In the upper two quartiles, pensions account for a much larger percentage than they did in the bottom two quartiles, but private pensions still accounted for only 12 percent of the income for those in both of the upper two quartiles and public pensions accounted for 7 percent for those in the third quartile and 13 percent for those in the fourth (highest) quartile. Examining the income quartiles by the ages of the elderly shows that for the individuals ages 65–69, the distribution of income sources across the income quartile groups is very similar to that of the entire elderly population (Figure 11). Pension income accounts for no more than 5 percent of the income in the two lower income quartiles, but increases for the upper two quartiles. However, for the 65–69 year olds, pension income does not account for as much of the income in the upper two quartiles relative to the entire elderly population, as private pensions only account for 9 percent and public pensions for 10 percent of the income among those in the highest quartile. For those ages 85 or older, Social Security overwhelmingly dominates as the source of the income for those in the lower three quartiles, as 91 percent, 88 percent, and 68 percent (lowest quartile to third quartile) of income was from Social Security (Figure 12). However, beginning in the

0%

10%

20%

30%

40%

50%

26%

9%

65-69

10%

19%

5%

OASDI

12% 9%

70-74

8% 5%

7%

Private Pension

25%

5%

Public Pension

Age

75-79

8% 8%

27%

4%

Source: Employee Benefit Research Institute estimates from the 1988 March Current Population Survey.

Earnings

31%

41%

46%

47%

8% 5%

Investment Income

80-84

6%

30%

Figure 8 Percentage of the Elderly's Income From Major Sources, by Age, 1987

2%

5% 6%

30%

Other

85 or older

50%

7%

Employee Benefit Research Institute 23

0%

10%

20%

30%

40%

50%

60%

11%

Earnings

65-69

9% 9%

31%

3%

OASDI

24%

3%

15%

Private Pension

12% 9%

70-74

12%

40%

3%

Public Pension

Age

75-79

14% 11% 11%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

37%

47%

7%

5%

Investment Income

80-84

14% 12% 11%

51%

Figure 9 Percentage of the Elderly's Income From Major Sources, by Age, 2004

4%

Other

85 or older

14% 11% 9%

58%

4%

24 Retirement Security in the United States

Income Quartile

0%

3%

9%

Earnings

10%

23%

20%

34%

OASDI

30%

41%

60%

12%

Public Pension

50%

92%

84%

Private Pension

40%

60%

21%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

All

Lowest 25%

25%-50%

50%-75%

Highest 25%

Figure 10 Percentage of the Elderly's Income From Various Sources, by Income Quartile, 2004

80%

9%

Investment Income

70%

10%

12%

13%

7%

8%

4%

4%

Other

90%

12%

100%

3%

1% 2% 2% 3%

3% 2% 5% 3%

16%

Employee Benefit Research Institute 25

Income Quartile

1%

7%

Earnings

10%

18%

OASDI

20%

37%

47%

40%

Private Pension

30%

31%

60%

Public Pension

50%

93%

79%

54%

18%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

All

Lowest 25%

25%-50%

50%-75%

Highest 25%

9%

9%

80%

11%

10%

Investment Income

70%

9%

Figure 11 Percentage of Americans' Income From Various Sources, Ages 64–69, by Income Quartile, 2004

7%

2%

3%

3%

11%

Other

100%

3%

1% 2% 2% 2%

2% 7%

90%

3%

7%

13%

26 Retirement Security in the United States

Income Quartile

4%

0%

1%

2%

7%

Earnings

10%

OASDI

20%

29%

40%

Private Pension

30%

58%

68%

60%

18%

Public Pension

50%

91%

88%

15%

Source: Employee Benefit Research Institute estimates from the 2005 March Current Population Survey.

0%

All

Lowest 25%

25%-50%

50%-75%

Highest 25%

4%

3%

3%

6%

Other

14%

100%

4%

1% 4% 2% 3%

1%

9%

3%

Investment Income

6%

90%

9%

12%

25%

80%

70%

11%

Figure 12 Percentage of Americans' Income From Various Sources, Ages 85 or Older, by Income Quartile, 2004

Employee Benefit Research Institute 27

28

Retirement Security in the United States

third quartile, pensions became a significant source of income as 12 percent and 15 percent of the income among those in the third and fourth quartiles came from private pensions, while 6 percent and 18 percent was from public pensions. These levels in the fourth quartile are significantly higher than of those for the 65–69-year-olds in that same quartile. Furthermore, among those ages 85 and older, investment income for those in the fourth quartile becomes a significantly more important source than for those in other quartiles and for those younger in the same quartile, reaching 25 percent. Mean and Median Pension Levels On average, the elderly had what many would consider rather modest incomes in 2004: the mean (average) annual income was $22,918 and the median (midpoint, half above and half below) annual income was $14,650 (Figure 13). Data show that the amount of income the elderly received decreased as their age increased. Those ages 65–69 had a mean income of $28,314 in 2004, compared with $17,385 for those 85 or older. The average income for the third income quartile was $19,487; only those in the highest (fourth) income quartile had what many would consider sizable income: $55,790. Given the modest levels of the elderly’s income, it is not surprising that the mean and median levels of pension income for those receiving a pension are rather modest. The mean pension amount among those receiving a pension was $10,354, and the median was $6,636. The mean and median public-pension amounts were significantly higher, at $19,978 and $15,600, respectively. The mean level of private-pension income decreased as age increased, with those ages 65–69 having a mean income of $12,490, compared with $7,892 for those ages 85 or older (mean amounts by age for public pensions followed the same pattern). The mean private-pension amount increased with income quartile, reaching $6,265 for the third quartile, before making a significant jump for those in the fourth quartile to $18,309. These relative results for the elderly’s income have persisted over time. The mean level of pension income grew for those receiving it from 1987 to 2004, so that the younger elderly did have higher pension incomes. However, there was a decrease in private pension recipiency, with the share of total income from pensions also declining in the most recent years.

7,521

80-84 10,616

4,601

1,987

954

3,516

3,408

4,105

4,448

8,250

4,301

1,583

600

2,200

2,400

2,957

3,048

4,200

Mean

16,395

7,871

3,579

1,672

6,471

7,926

8,339

10,077

10,253

$9,403

14,000

7,462

3,000

1,260

3,808

6,062

6,300

8,045

8,400

$7,200

Median

Mean

Median

Mean

48,839

18,066

9,772

4,515

12,572

13,266

16,264

15,815

18,362

39,181

17,406

9,685

5,112

9,194

9,973

11,124

11,300

12,128

14,957

5,723

2,317

1,390

5,164

4,142

5,852

6,989

8,001

$16,061 $10,961 $6,640

12,000

5,184

1,800

1,068

3,300

2,760

3,600

4,800

5,400

$4,296

Median

Mean

Note: All estimates are in nominal dollars.

Median

9,960 20,397

22,999

3,816

1,812

7,200

9,911

10,296

10,800

11,772

$10,500

10,062

4,309

2,079

9,868

12,959

13,937

13,549

13,984

$13,440

Source: Employee Benefit Research Institute estimates from the 1988, 1995, and 2005 March Current Population Survey.

30,776 10,046

14,706

50%-75% 15,222

Highest 25% 37,898

7,608

7,838

25%-50%

3,866

3,544

Lowest 25%

Income Quartile

6,788

7,965

75-79 11,156

9,279

9,160

70-74 12,926

85 or older

9,615

65-69 13,755

5,721

$3,264

Age

Median

Median

Mean

Mean

Public Pensions

Private Pensions

Total Income

Public Pensions

Private Pensions

All $12,320 $8,543 $4,715

Total Income

Of those with

1994

Of those with

Of those with

Of those with

1987

Median

2004

Mean

55,790

19,487

11,537

4,892

17,385

19,057

20,411

23,703

28,314

41,605

19,017

11,599

5,815

13,411

13,839

13,835

14,799

17,307

18,309

6,265

2,705

2,464

7,892

8,686

8,982

11,745

12,490

15,000

5,808

2,100

1,740

4,800

5,388

6,000

7,200

9,000

$6,636

Median

Private Pensions

Of those with

$22,918 $14,650 $10,354

Mean

Total Income

Figure 13 Elderly Americans' Mean and Median Annual Income, Mean and Median Annual Pension Income for those Having Such Pensions, by Age and Income Quartile, 1987, 1994, and 2004

Of those with Median

27,147

11,098

5,794

2,979

17,394

18,042

19,952

19,742

22,095

24,000

10,800

5,364

2,875

13,200

13,200

15,600

16,848

19,200

$19,978 $15,600

Mean

Public Pensions

Employee Benefit Research Institute 29

30

Retirement Security in the United States

Current Status of Retirees While the previous material in this chapter documented the incidence and average amounts of retirement income by source, the real question from a public policy perspective is whether current and future retirees will be able to afford an adequate standard of living in retirement. Significant work has been done by EBRI and others to evaluate how much workers will need in order to have the same after-tax and aftersavings amount for consumption in retirement that they enjoyed prior to retirement.2 Although this may be desirable from the standpoint of financial planning, it sets a goal that may be unrealistically high for many segments of the population. Another standard is used in this chapter to assess the current state of the retirement system. Instead of attempting to determine what percentage of the population will be able to attain a specified replacement ratio,3 this analysis attempts to model what percentage of retirees will have sufficient retirement wealth to pay for a basket of non-luxury goods in retirement for the remainder of their simulated life-paths.4 The expenditures used in the model for the elderly consist of two components—deterministic (unchanging) and stochastic (variable) expense assumptions. The deterministic expenses include those expenses that the elderly incur from a basic need or want of daily life, while the stochastic expenses in this model are exclusively health-event related—e.g., an admission to a nursing home or the commencement of an episode of home health care—that occur only for a portion, if ever, during retirement, not on an annual basis. Deterministic Retirement Expense Assumptions The deterministic expenses are broken down into seven categories—food, apparel and services (dry cleaning, haircuts), transportation, entertainment, reading and education, housing, and basic health expenditures. Each of these expenses is estimated for the elderly (age 65 or older) by family size (single or couple) and family income (less than $15,000, $15,000 to $29,999, and $30,000, or more in 2002 dollars) of the family/individual. The estimates are derived from the 2000 Consumer Expenditure Survey (CES) conducted by the Bureau of Labor Statistics of the U.S. Department of Labor. The survey targets the total noninstitutionalized population (urban and rural) of the United States and is the basic source of data for revising the items and weights in the market basket of consumer purchases to be priced for the Consumer Price Index. CES data

Employee Benefit Research Institute

provide detailed data on expenditures and income of consumers, as well as the demographic characteristics of those consumers. The survey does not provide state estimates, but it does provide regional estimates. Thus, the estimates are broken down into four regions—Northeast, Midwest, South, and West—to account for the differences in the cost of living across various parts of the country.5 Consequently, an expense value is calculated using actual experience of the elderly for each region, family size, and income level by averaging the observed expenses for the elderly within each category meeting the above criteria. The housing expenses are further broken down by whether the elderly own or rent their home. The basic health expenditure category has additional data needs beyond the CES. The basic health expenditures are estimated using a somewhat different technique and are comprised of two parts. The first part uses the CES as above to estimate the elderly’s annual health expenditures that are paid out-of-pocket or are not reimbursed (covered) or at least not fully reimbursed by Medicare and/or private Medigap health insurance, e.g., prescription drugs. The second part contains insurance premium estimates, including Medicare Part B premiums, and is not income related. All of the elderly are assumed to participate in Part B, and the premium is determined annually by the Medicare program and is the same nationally. For the Medigap insurance premium, all of the elderly are assumed to purchase a Medigap policy. A regional estimate is derived from a 2000 survey done by Weiss Ratings Inc., that received average quotes for three popular types of Medigap policies (A, F, and J) in 47 states and the District of Columbia. The estimates are calculated from the three policy types averaged over the states in the respective regions to arrive at the estimate for each region. This approach is taken for two reasons. First, sufficient quality data do not exist for the matching of retiree medical care (as well as the generosity of and cost of the coverage) and Medigap policy use to various characteristics of the elderly. Second, the health status of the elderly at the age of 65 is not known, let alone what it will be over the entire course of their remaining lives. Thus, assuming that everyone one has a standard level of coverage eliminates the need to differentiate among all possible coverage types as well as determining whether the sick or healthy have the coverage. Therefore, averaging the expenses over the entire population should have offsetting effects in the aggregate.

31

32

Retirement Security in the United States

The total deterministic expenses for elderly individual or family are then the sum of the value in all the expense categories for family size, family income level, and region of the individual or family. These expenses make up the basic annual (recurring) expenses for the individual or family. However, if the individual or family meet the income and asset tests for Medicaid, Medicaid is assumed to cover the basic health care expenses (both parts), not those of the individual or family. Furthermore, Part B premium relief for the low-income elderly (not qualifying for Medicaid) is also incorporated. Stochastic Retirement Expenditures The second component of health expenditures is the result of simulated health events that would require long-term care in a nursing home or home-based setting for the elderly. Neither of these simulated types of care would be reimbursed by Medicare because they would be for custodial (not rehabilitative) care. The incidence of the nursing home and home health care and the resulting expenditures on the care are estimated from the 1999 National Nursing Home Survey (NNHS) and the 2000 National Home and Hospice Care Survey (NHHCS). NNHS is a nationwide sample survey of nursing homes, their current residents and discharges that was conducted by the National Center for Health Statistics from July through December 1999. The NHHCS is a nationwide sample survey of home health and hospice care agencies, their current and discharge patients that was conducted by the National Center for Health Statistics from August through December 2000. For determining whether an individual has these expenses, the following process is undertaken. An individual reaching the Social Security normal retirement age has a probability of being in one of four possible assumed “health” conditions, based on the estimates of the use of each type of care from the surveys above and mortality: 

Not receiving either home health or nursing home care.



Home health care patient.



Nursing home care patient.



Death.

The individual is randomly assigned to each of these four categories with the likelihood of falling into one of the four categories based upon the estimated probabilities of each event. If the individual does not need

Employee Benefit Research Institute

long-term care, no stochastic expenses are incurred. Each year, the individual will again face these probabilities (the probabilities of being in the different conditions will change as the individual becomes older after reaching age 75 then again at age 85). This continues until death or the need for long-term care. For those who have a resulting status of home health care or nursing home care, the duration of care is simulated based on the distribution of the durations of care found in the NNHS and NHHCS. After the duration of care for a nursing home stay or episode of home health care, the individual will have a probability of being discharged to one of the other three conditions based on the discharge estimates from NNHS and NHHCS, respectively. The stochastic expenses incurred is then determined by the length of the stay/number of days of care times the per diem charge estimated for the nursing home care and home health care, respectively, in each region. For any person without the need for long-term care, this process repeats annually. The process repeats for individuals receiving home health care or nursing home care at the end of their duration of stay/care and subsequently if not receiving the specialized care again at their next birthday. Those who are simulated to die, of course, are not further simulated. As with the basic health care expenses, the qualification of Medicaid by income and asset levels is considered to see how much of the stochastic expenses must be covered by the individual to determine the individual’s final expenditures for the care. Only those expenditures attributable to the individual—not the Medicaid program—are considered as expenses to the individual and as a result in any of the “deficit” calculations. Total Retirement Expenses The elderly individual or families’ expenses are then the sum of their assumed deterministic expenses based upon their demographic characteristics plus any simulated stochastic expenses that they may have incurred. In each subsequent year of life, the total expenditures are again calculated in this manner. The base year’s expenditure value estimates excluding the health care expenses are adjusted annually using the assumed general inflation rate of 3.3 percent from the 2001 OASDI Trustees Report, while the health care expenses are adjusted annually using the 4.0 percent medical consumer price index that corresponds to the June 2002-June 2003 level.6,7

33

34

Retirement Security in the United States

Comparison of Retirement Income and Retirement Expenses The primary objective of this analysis is to combine the simulated retirement income and wealth8 with the simulated retiree expenditures to determine how much each family unit would need to save today (as percentage of their current wages) to maintain a prespecified “comfort level” (i.e., confidence level) that they will be able to able to afford the simulated expenses for the remainder of the lifetime of the family unit (i.e., death of second spouse in a family). These savings rates are reported by age cohort, family status (at retirement) and gender. Six five-year birth cohorts are simulated. The oldest group was born in the period 1936 to 1940 inclusive (currently age 66 to 70), while the youngest group was born in the period 1961 to 1965 (currently age to 41 to 45), inclusive. Three combinations of gender/family status at retirement were reported: family, single male, and single female. In addition, the relative income was reported by estimating lifetime income quartiles (from 2002 though retirement age) for each of the 18 combinations of birth cohort and gender/family status at retirement. It is important to note that within each of the groups modeled there will undoubtedly be significant percentages in the zero category, as well as those at levels far higher than most individuals could possibly save. These situations are accounted for in two ways: First, Medians are reported for each of the groups; in other words, the numbers presented in Figures 14–15 provide a number representing the estimate for the 50th percentile when ranked by percentage of compensation. Second, the reported values are limited to 25 percent of compensation under the assumption that few, if any, family units would be able to contribute in excess of this percentage on a continuous basis until retirement age. It is also important to note that these percentages merely represent savings that need to be generated in addition to what retirement income and/or wealth is simulated by the model. Therefore, if the family unit is already generating savings for retirement beyond what is included in defined benefit or defined contribution plans, IRAs, Social Security and/ or net housing equity, that value needs to be deducted from the estimated percentages. After the retirement income and wealth was simulated for each family unit, 1,000 observations were simulated (from retirement age until death of the individual for single males and single females or the second person to die for families), and the present value of the aggregated deficits at retirement age were computed. At that point, the observations were rank-

Employee Benefit Research Institute

ordered in terms of the present value of the deficits, and the 75th and 90th percentiles of the distribution were determined. Next, the future simulated retirement income accumulated to retirement age was determined, and the information used to determine the percentage of compensation that would need to be saved to have sufficient additional income to offset the present value of accumulated deficits for the 75th and 90th percentiles of the distribution. Results Figure 14 shows the median percentage of compensation that must be saved each year until retirement for a 75-percent confidence level when combined with simulated retirement wealth, assuming current Social Security benefits and that housing equity is never liquidated. For example, all gender/family combinations in the first two income quartiles for the oldest birth cohort are at the 25 percent of compensation threshold. For those in the highest income quartile for this birth cohort, the percentages of additional annual compensation needed to be saved are 23.8 percent for singe females, 13.9 percent for single males, and 6.1 percent for families. Figure 15 shows the additional savings required to provide retirement adequacy in 9 out of 10 simulated life paths. In this case, nearly all of the gender/family status-at-retirement combinations for the first three income quartiles are at the threshold (the median for families in the third quartile is estimated at 24.8 percent of compensation). Those in the highest income quartile for this birth cohort all have requirements that would prove difficult if not impossible to implement: Single females are estimated to now need to save more than 25 percent of compensation, single males 22.1 percent of compensation, and families 10.1 percent of compensation. Given that most individuals would be unlikely to choose a situation that would provide them with adequate retirement income only 50 percent of the time, this analysis focuses only on the 75-percent and 90-percent confidence levels.9 Figure 16 provides another way of illustrating which cohorts may be the most vulnerable to inadequate financial resources in retirement. This figure starts with the baseline scenario described above (current Social Security benefits and no liquidation of net housing equity) and assumes that each worker contributes an additional 5 percent of compensation from 2003 until retirement age to supplement his or her Social Security and tax-qualified retirement plans. The percentage of each cohort estimated to have sufficient retirement income and/or wealth to cover

35

1

2

3

1936–1940

4

1

2

3

1941–1945

Source: EBRI-ERF Retirement Security Projection Model.

0%

5%

10%

15%

20%

a 25%

4

1

3

1

3

1951–1955

2

single male

Birth Cohort/Income Quartiles

4

single female

1946–1950

2

family

4

1

3

1956–1960

2

Figure 14 Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 75% Chance of Covering Basic Retirement Expenses (assumes current Social Security and housing equity is never liquidated)

4

1

a

2

3

4

25% = 25% or more.

1961–1965

36 Retirement Security in the United States

1

3

1936–1940

2

4

1

3

1941–1945

2

Source: EBRI-ERF Retirement Security Projection Model.

0%

5%

10%

15%

20%

25%

a

4

1

3

1

3

1951–1955

2

single male

Birth Cohort/Income Quartile

4

single female

1946–1950

2

family

4

1

3

1956–1960

2

Figure 15 Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 90% Chance of Covering Basic Retirement Expenses (assumes current Social Security and housing equity is never liquidated)

4

a

3

1961–1965

2

25% = 25% or more.

1

4

Employee Benefit Research Institute 37

1

3

1936–1940

2

4

1

3

1941–1945

2

4

1

3

4

1

3

1951–1955

2

single male

Birth Cohort/Income Quartile

1946–1950

2

single female

4

1

3

1956–1960

2

4

1

3

1961–1965

2

4

Source: EBRI-ERF Retirement Security Projection Model. Assumes current Social Security, and that housing equity is never liquidated. The model includes the possibility of chronic long-term home health care and nursing home expenses.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

family

Figure 16 Percentage of Retirees Estimated to Have Sufficient Retirement Income/Wealth by Saving 5% of Compensation Each Year From 2003 Until Retirement (assumes current Social Security benefits)

38 Retirement Security in the United States

Employee Benefit Research Institute

the simulated retirement expenses described earlier is displayed. For example, approximately 30 percent of the simulated life-paths for the lowest income quartile for those in the 1936-1940 birth cohort would be expected to have sufficient retirement resources. However, at least 85 percent of the simulated life-paths for the third or fourth income quartiles for those in the 1961-1965 birth cohort would be sufficient. This is in large part due to the fact that the younger cohorts will have additional years to accumulate the additional 5 percent of compensation. For each birth cohort, the lower-income quartiles have more risk of insufficient retirement income than their higher-paid counterparts. Moreover, single females tend to exhibit more vulnerability than single males, while families are typically the least vulnerable.

39

III

Changes and Factors Affecting Future Retirement Income Prospects

41

T

he United States and its workers are facing many changes that will lead to challenges for taxpayers, employers, and retirees to fund their retirement expenditures, particularly if retirees hope to maintain their standard of living throughout their retirement years. The aging of the United States population, increased global competition, and rapidly increasing health care costs are the leading causes of the challenges, as the large post-World War II baby boom generation reaches retirement age.

Aging of United States Population In 1987, 28.2 percent of the American workforce was age 25–34 (Figure 17). By 2004, this had declined to 21.3 percent. In contrast, 15.0 percent of workers were age 45–54 in 1987, which grew to 22.7 percent by 2004. Furthermore, 20.4 percent of workers in 1987 were under age 25, compared with just over 15 percent in 2004, while the percentage of workers age 55 or older increased from 13.5 percent in 1987 to 17.1 percent in 2004.

Figure 17 Age Distribution of the American Workforce, 1987–2004 Age 20 or younger

1987

1992

1997

2002

2004

10.3%

8.3%

8.6%

7.5%

7.1%

21-24

10.1

9.4

8.1

8.2

8.3

25-34

28.2

26.5

23.6

21.7

21.3

35-44

22.9

25.7

26.4

24.6

23.7

45-54

15.0

17.4

20.0

22.2

22.7

55-64

10.1

9.5

10.0

12.1

12.9

65 or older

3.4

3.3

3.4

3.8

4.2

Source: Employee Benefit Research Institute estimates of the 1988, 1993, 1998, 2003, and 2005 March Current Population Survey.

Changes and Factors Affecting Future Retirement Income Prospects

43

44

Retirement Security in the United States

As this large group of workers becomes retirees, there will be significantly fewer workers in proportion to retirees relative to prior years—a crucial factor affecting the financing of the Social Security program. In 1965, there were 4.0 Social Security-covered workers for every Social Security beneficiary (Figure 18). In 2005, this number had fallen to 3.3 workers per beneficiary. Once all of those in the baby boom generation reach age 65 in about 2030, there will be just over two workers for every beneficiary. This ratio is then projected to fall below two by 2065.

Retiree Health Care At a time when a greater percentage of the American workforce is aging, the costs incurred by private-sector employers that provide health care for retirees has been increasing by double-digit growth rates. In 2002, the average increase in total retiree health care costs for privatesector firms with 1,000 or more employees increased by 16.7 percent (Figure 19). While these growth rates have slowed through 2005, the rate in 2005 was still 10.3 percent. At the same time these costs increased, the percentage of mid- to large-sized firms (with 200 or more employees) offering retiree health benefits have declined from 40 percent in 1999 to 35 percent in 2005 (Figure 20).

Retirement Plan Participation The percentage of American workers participating in an employment-based retirement plan (either a defined benefit or defined contribution plan) has been fairly constant, increasing slightly from 1987 to 2004. Among all workers, the percentage who participated in a plan increased from 37.6 percent in 1987 to 41.9 percent in 2004, with even higher levels from 1998 to 2001 (Figure 21). When focusing more specifically on private-sector wage and salary workers (excluding self-employed workers) ages 21–64, the participation rate increased from 39.8 percent in 1987 to 43.0 percent in 2004. Retirement plan rates of participation of full-time, full-year workers and public-sector workers were higher but unchanged from 1987 to 2004. While the percentage of workers who participate in an employmentbased retirement plan has remained relatively constant, the type of plan that private-sector workers are participating in has changed dramatically. In 1979, 62 percent of private-sector workers who were participating in an employment-based retirement plan had a defined benefit (pension) plan solely, with another 22 percent having a defined benefit plan along with a defined contribution (401(k)-type) plan (Figure 22). Only 16

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1975

3.2

1985

3.3

1995

3.3

2005

3.3

2015

2.9

2025

2.3

2035

2.1

2045

2.0

Source: 2005 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, http://www.ssa.gov/OACT/TR/TR05/index.html.

1965

4.0

Figure 18 Ratio of Old-Age, Survivor, and Disability Insurance (OASDI)Covered Workers Per OASDI Beneficiary, 1965–2075

2055

2.0

2065

1.9

2075

1.9

Employee Benefit Research Institute 45

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

2003

13.7%

Source: Kaiser/Hewitt 2002-2005 Survey on Retiree Health Benefits, December 2002-2005

2002

16.0%

2004

12.7%

Figure 19 Average Increase in Total Retiree Health Costs for PrivateSector Firms With 1,000 or More Employees, 2002–2004

2005

10.3%

46 Retirement Security in the United States

1988

66%

1991

46%

1993

36%

1995

40%

1998

40%

1999

40%

2000

35%

2001

37%

2002

36%

2003

38%

2004

36%

2005

35%

Source: Kaiser/HRET Survey of Employer Sponosered Health Benefits, 1999-2005; KPMG Survey of Employer-Sponsored Health Benefits, 1991, 1993, 1995, 1998; The Health Insurance Association of America (HIAA), 1988. Figure available at www.kff.org/insurance/7315/sections/ehbs05-11-1.cfm.

0%

10%

20%

30%

40%

50%

60%

70%

Figure 20 Percentage of All Large Firms (200 or More Workers) That Offer Health Benefits Also Offering Retiree Health Benefits, 1988–2005

Employee Benefit Research Institute 47

35%

40%

45%

50%

55%

60%

65%

70%

75%

80%

1988

38.0%

40.0%

46.2%

58.3%

75.6%

1993

39.0%

40.1%

46.2%

57.7%

74.5%

1994

41.0%

42.1%

48.2%

59.6%

77.6%

1995

41.0%

42.2%

47.8%

57.9%

76.7%

1996

41.7%

43.1%

48.4%

59.0%

77.1%

1997

41.8%

44.2%

49.1%

59.2%

75.5%

1998

43.6%

46.1%

51.0%

60.3%

77.2%

1999

43.8%

46.1%

51.1%

60.4%

77.2%

2000

44.4%

46.7%

51.6%

59.8%

77.3%

2001

43.0%

45.1%

49.8%

58.3%

75.3%

Public Sector Wage and Salary Workers Ages 21-64

2002

41.8%

43.1%

48.2%

56.7%

74.8%

Private Sector Wage and Salary Workers Ages 21-64

1992

39.4%

40.4%

46.5%

59.0%

75.2%

Full-Time, Full-Year Wage and Salary Workers Ages 21-64

1991

39.7%

40.8%

46.8%

58.9%

76.1%

All Wage and Salary Workers Ages 21-64

1990

39.0%

41.3%

47.4%

59.5%

76.1%

All Workers

1989

38.9%

40.9%

47.2%

59.0%

76.3%

Source: Employee Benefit Research Institute estimates from the 1987-2005 March Current Population Surveys.

1987

37.6%

39.8%

46.1%

58.4%

75.6%

Figure 21 Percentage of Various Work Forces Who Participate in an Employment-Based Retirement Plan, 1987–2004

2003

42.0%

43.2%

48.4%

57.1%

75.8%

2004

41.9%

43.0%

48.3%

56.6%

75.8%

48 Retirement Security in the United States

0%

10%

20%

30%

40%

50%

60%

70%

Defined Contribution Only

Both Plans

Source: Depart of Labor Form 5500 Tabulations 1999, 2000-2004 EBRI estimates.

Year

1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Defined Benefit Only

Figure 22 Distribution of Private-Sector Workers Who Are Participating in an Employment-Based Retirement Plan, by Plan Type, 1979–2004

Employee Benefit Research Institute 49

50

Retirement Security in the United States

percent of the participants had a defined contribution plan solely. However, by 2004, the percentages with only one plan type had switched, as 63 percent of the participants had a defined contribution plan solely and 10 percent had a defined benefit plan solely. Those participants with both types increased slightly, to 27 percent. Furthermore, the method for determining benefits within the socalled “traditional” pension (defined benefit) plans has also been changing during this transformation toward defined contribution plans. Most defined benefit plans have traditionally used a career average pay or finalaverage pay method to determine benefits. However, other arrangements have become more common among defined benefit plans, as 5 percent of defined benefit plans (and more significantly, 25 percent of participants) were in so-called “hybrid” defined benefit plans—most commonly a cash balance plan—in 2003 (Figure 23). Along with the change in method of benefit determination, the allowable payout methods also changed. The percentage of participants in a defined benefit that had a lump-sum option increased from 15 percent in 1995 to 48 percent in 2002 (Figure 24). The net result has been a significant reduction in the number of remaining defined benefit pension plans that pay out a benefit in the form of a lifetime annuity.

Individual Choices Affecting Retirement Income Workers who are eligible to participate in a retirement plan have various choices to make that will fundamentally affect their ultimate level of retirement income. They have to choose whether to participate, decide how much to contribute when participating, how to invest their assets while contributing, and then what to do with any accumulated benefits when changing jobs or retiring, among other decisions. Participation Rates In 2001, 75 percent of defined contribution-eligible participants participated in an offered plan (Figure 25). This was higher than the 74 percent level in 1995, but below the peak of 77 percent in 1998. The participation rate is strongly affected by the income of the individual, with the likelihood of participation increasing with income. Specifically, in 2001, 53 percent of eligible workers with family incomes below $10,000 participated in the offered defined contribution plan, compared with 85 percent of those with family incomes of $100,000 or more.

0%

5%

10%

15%

20%

25%

30%

2001

21%

4%

Year

2002

23%

Source: Pension Benefit Guaranty Corporation, Pension Insurance Data Book, 2004. (www.pbgc.gov/docs/2004databook.pdf).

4%

Participants

Plans

5%

Figure 23 Percentage of Private-Sector Defined Benefit Plans Who Are Hybrid Plans and Participants That Are in Hybrid Plans, 2001–2003

2003

25%

Employee Benefit Research Institute 51

0%

10%

20%

30%

40%

50%

60%

1997

Source: U.S. Department of Labor, Bureau of Labor Statistics. National Compensaton Survey: Employee Benefits in Private Industry, 2002–2003. www.bls.gov/ncs/ebs/sp/ebbl0020.pdf

1995

15%

23%

Figure 24 Percentage of Private-Sector Defined Benefit Plan Participants Having a Lump-Sum Distribution Option Available, 1995, 1997, and 2002

2002

48%

52 Retirement Security in the United States

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

All

77%

75%

42%

less than $10,000

44%

53%

1995

1998

54%

61%

75% 69%

$25,000-$49,999

72%

Family Income

$10,000-$24,999

56%

2001

Source: Employee Benefit Research Institute estimates of the 1995, 1998, and 2001 Survey of Consumer Finances.

74%

78%

$50,000-$99,999

79%

84%

Figure 25 Participation Rates of Family Heads Eligible for an EmploymentBased Defined Contribution Plan, 1995, 1998, and 2001 86% 85%

$100,000 or more

83%

Employee Benefit Research Institute 53

54

Retirement Security in the United States

Contribution Rates In 2003, the average contribution to salary reduction plans (i.e., 401(k), 403(b), or thrift savings plans) was 7.5 percent of annual earnings (Figure 26). This is an increase from 6.6 percent to such plans in 1988. Asset Allocation and Account Balances10 In 2004, 46 percent of total 401(k) plan assets were invested in equities, 10 percent in balanced funds, 15 percent in company stock, 10 percent in bond funds, 4 percent in money funds, and 12 percent in guaranteed investment contracts and stable value funds. These numbers have essentially been at these levels since 1996. Furthermore, as the age of the participant increases the percent in equities declines (50.9 percent for those in their 40s compared with 36.5 percent for those in their 60s). The average 401(k) plan balance was found to $91,042 in 2004 for those in the study from 1999 through 2004, which was an increase from $67,016 in 1999. However, the balance varied significantly by the age and tenure of the participants, with the oldest and longest tenure participants having an average balance of approximately $180,000. Lump-Sum Distributions The percentage of individuals who took a lump-sum distribution from an employment-based retirement plan and rolled over the entire amount of the distribution to a tax-qualified savings vehicle (i.e, individual retirement accounts, another job’s plan) has increased with time before leveling off. Among those distributions taken before 1980, 21 percent were entirely rolled over (Figure 27). This percentage increased to 48 percent for those distributions taken from 1994–1998, before falling to 45 percent for distributions taken from 1999–2003. However, the percentage of individuals who have taken a lump-sum distribution and used any proportion for consumption has continued to decrease, from 36.4 percent of distributions taken before 1980 to 23.4 percent for those taken from 1999–2003 (Figure 28). The amount of the distribution affects the likelihood of an individual rolling over the distribution: The larger the amount, the more likely it will be rolled over. For distributions of $1–$499 (2003 dollars), 19.9 percent of lump-sum recipients rolled over their entire distribution (Figure 29). This increased with the amount of distribution to 68.1 percent for distributions of $50,000 or more.

Percentage of Earnings

6.0%

6.2%

6.4%

6.6%

6.8%

7.0%

7.2%

7.4%

7.6%

7.8%

1993

Year

1998

2003

7.5%

Source: Employee Benefit Research Institute estimates of the May 1988 and April 1993 Current Population Survey employee benefit supplements and the 1996 and 2001 Panel of the Survey of Income and Program Participation Topical Module 7.

1988

6.6%

7.1%

7.4%

Figure 26 Average Percentage of Annual Earnings Contributed to a Salary Reduction Plan, Civilian Nonagricultural Wage and Salary Workers, Age 16 and Over, 1988, 1993, 1998, and 2003

Employee Benefit Research Institute 55

Before 1980

21%

1980-1986

31%

Year of Distribution

1987-1993

46%

1994-1998

48%

a

Includes investment in individual retirement accounts (IRAs), rollovers to IRAs, individual annuities, and other employment-based retirement plans.

Source: Employee Benefit Research Institute estimates from 2001 Panel of the Survey of Income and Program Participation Topical Module 7.

0%

10%

20%

30%

40%

50%

60%

1999-2003

45%

Figure 27 Proportion of Lump-Sum Recipients Using Entire Portion of Their Most Recent Distribution a for Tax-Qualified Financial Savings, by Year of Receipt, Civilians Age 21 and Over

56 Retirement Security in the United States

Before 1980

36.4%

1980-1986

30.9%

Year of Distribution

1987-1993

24.7%

1994-1998

24.3%

a

Includes purchases of consumer items (car, boat), medical and dental expenses, general everyday expenses, and other spending.

Source: Employee Benefit Research Institute estimates from 2001 Panel of the Survey of Income and Program Participation Topical Module 7.

0%

5%

10%

15%

20%

25%

30%

35%

40%

Figure 28 Proportion of Lump-Sum Recipients Reporting Using Any Portion of Their Most a Recent Distribution for Consumption, by Year of Receipt, Civilians Age 21 and Over

1999-2003

23.4%

Employee Benefit Research Institute 57

0%

10%

20%

30%

40%

50%

60%

70%

80%

21.3%

$500-$999

$1,000-$2,499

30.3% 30.4%

$5,000-$9,999

38.8%

Amount of Distribution

$2,500-$4,999

36.5% 35.4%

40.8%

$10,000-$19,999

45.0% 46.7%

a

59.2%

$20,000-$49,999

56.5%

Includes investment in individual retirement accounts (IRAs), rollovers to IRAs, individual annuities, and other employment-based retirement plans.

Source: Employee Benefit Research Institute estimates from 2001 Panel of the Survey of Income and Program Participation Topical Module 7.

$1-$499

19.9%

25.5% 24.5%

Distribution Year Dollars

2003 Dollars

68.1%

73.4%

$50,000 or more

Figure 29 Proportion of Lump-Sum Recipients Using Entire Portion of Their Most Recent Distribution for Tax-Qualified Financial Savings,a by the Amount of the Most Recent Distribution, Civilians Age 21 and Over

58 Retirement Security in the United States

Employee Benefit Research Institute

A job-changer or retiree also could have the choice to retain his or her retirement plan benefits in a previous employer’s plan. In fact, 38 percent of workers who were former employment-based plan participants retained their benefits through a previous employer’s plan in 2003 (Figure 30). Another 11 percent retained some of their benefits at a previous employer’s plan. Therefore, only 51 percent of those individuals who were participants in a prior employer’s plan either took a lump sum by choice or were forced to take it. This decision was also affected by the size of the distribution, as those with larger distributions were more likely to retain their benefits at a previous employer’s plan (Figure 31). Among distributions of $1–$499, 12.6 percent were retained at a previous employer, compared with 36.9 percent for distributions of $50,000 or more. Consequently, the percentage of individuals with large accrued benefits is much more likely to retain their benefits than what is shown in just the lump-sum distribution analysis, because those individuals who retain benefits with their previous employer are not included in those numbers. This means that approximately 80 percent of individuals who change jobs or retire, and have an account balance of $50,000 or more, are preserving their benefits for a later use (hopefully retirement).

59

Received a Lump Sum Distribution Only 51%

Source: Employee Benefit Research Institute estimates from the 2001 Survey of Income and Program Participation Topical Module 7.

Received a Lump Sum Distribution and Retained Retirement Plan Benefits at a Previous Job 11%

Retained Retirement Plan Benefits at a Previous Job Only 38%

Figure 30 Percentage of Former Participants in an Employment-Based Retirement Plan Who Either Took a Lump-Sum Distribution or Retained Their Benefits in Their Previous Job's Plan, 2003

60 Retirement Security in the United States

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

28.3%

$1-$499

12.6%

14.1% 14.0%

$1,000-$2,499

86.0%

18.5%

$2,500-$4,999

81.5%

34.7%

$5,000-$9,999

65.3%

39.7%

$10,000-$19,999

60.3%

Left in Plan

Lump Sum

Value of Most Recent Lump Sum Distribution or Current Plan Balance

$500-$999

85.9%

30.9%

$20,000-$49,999

69.1%

36.9%

$50,000 or more

63.1%

This only looks at those that took only a lump-sum distribution or retained their retirment plan benefits in a previous job's plan, not a combination of the two. Furthermore, only individuals that reported a balance in the previous job's plan were included. The balance is at the time of the survey, not at the time of the job termination, with all lumpsum distributions in 2003 $s.

a

Source: Employee Benefit Research Institute estimates from the 2001 Survey of Income and Program Participation Topical Module 7.

All

71.7%

87.4%

Figure 31 Percentage of Individuals Who Had a Retirement Plan at a Previous Job Who Took a Lump-Sum Distribution vs. Leaving the Assets in the Plan, by the Value of the Distribution and the Current Plan Balance, 2003a

Employee Benefit Research Institute 61

IV

Cash Balance Plans11

63

T

he trend in the late 1990s among large employers toward conversion of “traditional” pensions (final-average and career-average defined benefit plans) to cash balance plans has raised a controversial and complex set of issues. A cash balance plan is a “hybrid” type of pension plan—i.e., one that takes on the characteristics of both a defined benefit plan and a defined contribution plan. Legally, a cash balance plan is a defined benefit plan. A cash balance plan offers some of the popular advantages of a defined benefit plan but is designed to look more like a defined contribution12 plan, with an individual “hypothetical” account that appears to accumulate assets for each participant. Cash balance plan accounts are a record-keeping feature only, as these plans are funded on an actuarial basis, in the same way that defined benefit pension plans are funded. Therefore, at any point in time, the benefits promised to a participant are based on the plan formulae and not on the assets in his or her “account.” In a typical cash balance plan, a participant’s retirement account grows by earning annual credits that may be based on a flat percentage of pay but that might be integrated with Social Security benefits (Quick, 1999). Cash balance plans also provide a yield on the hypothetical account that is typically defined as either the 30-year Treasury rate or the one-year T-Bill rate plus a stated percentage (Gebhardtsbauer, 1999). Utilization of cash balance plans increased dramatically from 19951999.13 This trend was abruptly curtailed in the latter portion of the 1990’s due to litigation and an IRS moratorium on the issuance of determination letters for hybrid plans. Subsequent to the time this chapter was written, Congress enacted the Pension Protection Act of 2006, which resolved many of the uncertainties previously facing plan sponsors with respect to cash balance plans. The Act applies on a prospective basis and thus much of the following text will apply to plans converted prior to June 29, 2005.14

Cash Balance Plans

65

25

30

35

conversion at age 55 (without transition credit)

traditional final average defined benefit plan

cash balance (service weighted)

40

45

Age

50

55

60

Source: Statement of Jack VanDerhei before the Senate Health, Education, Labor and Pensions Committee, Hearing on Hybrid Pensions, Sept. 21, 1999 (EBRI Testimony T-121)

$0

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

Figure 32 Illustration of a Conversion From a Hypothetical Traditional Final-Average Defined Benefit Plan to a Hypothetical ServiceWeighted Cash Balance Plan (Without Transition Credits) at Age 55

65

66 Retirement Security in the United States

Employee Benefit Research Institute

Fundamental Economic Distinction Between Final-Average and Cash Balance Plans Under either the final-average or cash balance plans illustrated in Figure 32, an employee starting at age 25 will obtain the same benefit value at age 65 if he or she remains with the same employer for a full career. Nevertheless, the accrual rates under each plan differ fundamentally. The annual increase in benefit value (viz., how much additional retirement income an employee will earn by working one more year) tends to be much higher for young employees under the cash balance plan and much higher for older employees under the final-average plan. This is true even though the cash balance plan illustrated in this figure adopts a service-weighted pay credit schedule.15 A difference in accrual rates between older and younger workers on conversion from a final-average to a cash balance plan is likely to exist whether or not a so-called wear-away provision (explained later) is included in the plan. The difference is conceptually similar to the effects of changing a final-average plan to a career-average plan or, more drastically, terminating a defined benefit plan and establishing a defined contribution plan. However, the magnitude of the difference is influenced by plan-specific design parameters.16 Employees faced with the type of graph shown in Figure 32 are likely to wonder why the shapes look different. The difference essentially lies in the different determinants of benefit value under each type of plan. While the present value of the annual accrual of pension wealth expressed as a percentage of compensation under a final-average plan at any point in time depends on age, service, and pay, it depends predominantly on pay and service (and a lesser extent on age) under a cash balance plan. Therefore, even if the overall generosity of a plan remains the same after conversion to a cash-balance formula, higher accruals for young employees means that accruals for older employees will likely decrease unless some type of grandfathering or transition provisions (explained below) are provided to older workers. For example, an employee participating in the hypothetical finalaverage defined benefit plan in Figure 32 would have a present value from his or her defined benefit plan at age 55 of approximately $95,000, as opposed to approximately $135,000 for a similar employee who had participated in the hypothetical cash balance plan for the same period of 30 years. However, if the hypothetical final average plan were then converted to the hypothetical cash balance plan without the provision of

67

68

Retirement Security in the United States

any type of transition credit, the employee would not benefit from the rapid escalation in pension wealth from age 55 to 65 that is associated with the final average plan. Instead, during the final 10 years he or she would experience a slope of the accrual path similar17 to that experienced by the participant who remains under the cash balance plan for the entire 40 years. As a consequence, the participant will not end up with the same financial position at age 65 but, barring any transition provisions, would experience a decrease in pension wealth of approximately 23 percent. Another significant difference between a traditional defined benefit plan and a cash balance plan concerns the inherent uncertainty involved in estimating the nominal amount of retirement income. Traditional defined benefit plans are not typically thought of in this regard since the amount is specified in a formula and (with the exception of certain integrated plans) can be directly computed once the average compensation and years of participation are known. However, it appears that an increasing percentage of defined benefit participants are now receiving their distributions in the form of lump-sum distributions (LSDs)—a form that can provide great uncertainty to employees with respect to the amount that they will receive due to fluctuations in the relevant discount rates (Bone, 1999). In contrast, cash balance plans provide LSDs that are stabilized, but annuity values under these arrangements may be subject to fluctuations in annuity purchase prices (although it appears some employers are willing to hold annuity purchase rates constant in the plan) (Gebhartsbauer, 1999).

Potential Advantages: Cash Balance vs. Final-Average Plans Before discussing key public policy issues and the possible ramifications of modifying the existing legislative and/or regulatory landscape, it may be helpful to consider why a sponsor of a final-average defined benefit plan may be interested in converting to a cash balance plan:18 Ease of Communication vs. Invisible Plan Syndrome Sponsors of traditional defined benefit pension plans often bemoan the lack of recognition they receive from their employees, even though substantial sums of money are contributed and/or accrued annually. When the quality of workers’ information regarding traditional pension offerings was evaluated,19 about one-third of workers queried were unable to answer any questions about early retirement requirements, and about

Employee Benefit Research Institute

two-thirds of those who offered answers about early retirement were wrong (Mitchell, 1988). As an alternative to explaining the complex benefit formulas used by traditional defined benefit plans, conveying information through theoretical account balances under cash balance plans facilitates employee appreciation of both current pension wealth and the annual pay and interest credits that increase pension value over time. No Magic Numbers of Age and Service Final-average defined benefit plans often require employees to satisfy some combination of age and service before they are entitled to retire with an early retirement subsidy, and the magnitude of the dollar loss from leaving prior to that time can be substantial (Ippolito, 1998). In contrast, the accrual pattern under a cash balance plan typically does not have a sudden, rapid increase after attainment of specific age and service criteria. As a result, cash balance plans are more attractive to a mobile work force. Higher Benefits to Employees Who Do Not Stay With One Employer for Their Entire Career Figure 33 shows the percentage increases in annual retirement benefits at normal retirement age for an employee in a hypothetical cash balance plan versus a hypothetical final-average defined benefit plan. The figures in this figure are tabulated from a Congressional Research Service (CRS) report to Congress that includes calculations for two types of employees: (a) one who enters the employer’s plan at age 25 and remains in that plan for 40 years and (b) one who changes jobs every 10 years (Purcell, 1999). Comparing the two sets of bar graphs, one can see that for a hypothetical individual staying at the same job for his or her entire life, the cash balance plan provides a larger benefit after the first 10 and 20 years of service. But, by age 55, the final-average plan is slightly more valuable, and by retirement age the benefit derived from the final-average plan would be 30 percent larger than the cash balance benefit. However, this “one-job for life” scenario only applies to small percentage of the work force (Yakoboski, 1999). Employees are more likely to have four, if not more, jobs during their careers. The second set of bar graphs show that in those cases, the series of cash balance plan benefits dominate those accrued under the final-average plans at every age, and the final retirement benefits are approximately 40 percent larger.20

69

67%

35

39% 30%

45

-2%

55

Attained Age (Years) (entry age = 25)

44%

four jobs

one job

45%

-30% 65

40%

Source: Statement of Jack VanDerhei before the Senate Health, Education, Labor and Pensions Committee, Hearing on Hybrid Pensions, Sept. 21, 1999 (EBRI Testimony T-121)

-40%

-20%

0%

20%

40%

60%

80%

Figure 33 Hypothetical Percentage Increases in Annual Benefits at Normal Retirement Age. Cash Balance vs. Final-Average Plan: Impact of Job Tenure

70 Retirement Security in the United States

Employee Benefit Research Institute

Potential Advantages: Cash Balance vs. Defined Contribution Plans Of course, an employer that sponsors a final-average plan also has the alternative of terminating the existing defined benefit plan (assuming it is adequately funded) and setting up a defined contribution plan through which to provide benefits for future service. However, several considerations may make this option problematic: Ease of Conversion vs. New Plan Establishment Whereas a conversion from a final-average defined benefit plan to a cash balance plan requires only a plan amendment (Rappaport, Young, Levell, and Blalock, 1997), terminating the same plan and setting up a successor defined contribution plan may trigger a reversion excise tax of either 20 percent or 50 percent (Alderson and VanDerhei 1991). If the defined benefit plan was overfunded, the surplus in a conversion to a cash balance plan would be used to reduce future contributions (as it would under the traditional plan); if it was underfunded, the unfunded liability is amortized in the normal fashion (Warshawsky, 1997). Guarantee of Employee Participation The noncontributory nature of most (if not all) cash balance plans eliminates the problem of eligible employees who choose not to participate in a retirement plan or make de minimis contributions in a 401(k) arrangement. As a result, employees are guaranteed a benefit under a cash balance plan without needing to actively choose to participate in the plan, and the plan is protected from possible disqualification due insufficient participation among lower-paid workers. Retirement Pattern Predictability Investment risk is typically directly borne by employers under a cash balance plan and by workers under a defined contribution plan (see Auer 1999, however, for one notable exception). As a result, the employer is better able to predict retirement patterns under a cash balance plan, since retirement income will not be susceptible to market fluctuations. Under a defined contribution plan, employers may face unexpected increases in early retirements during a strong bull market and unexpected delays of retirement during a market correction (especially if it is prolonged).

71

72

Retirement Security in the United States

Retirement Benefit Predictability Since employers directly bear investment risk under cash balance plans, they need not worry about overly conservative worker-investors in a defined contribution plan. Holden and VanDerhei (2005) report that 40 percent of 401(k) participants in their 20s and 28 percent of those in their 30s hold no equity funds in their 401(k) account balances. Although approximately one-half of these individuals in each age cohort have some equity market exposure through company stock and/or balanced funds, a significant percentage of them may be subjecting themselves to expected rates of return too low to generate sufficient retirement income at normal retirement age. Funding Flexibility Finally, a cash balance plan may have more funding flexibility than a defined contribution plan, depending on the type of commitment made to employees. Although some profit-sharing plans provide for annual contributions that are entirely discretionary for the plan sponsor (Allen, Melone, Rosenbloom, and VanDerhei, 1997), a defined benefit plan is the only vehicle that will allow employees to continue their normal benefit accruals while employer contributions are reduced or even temporarily curtailed.

Potential Limitations of a Conversion From a Defined Benefit to a Cash Balance Plan Cash balance plans are appealing to employers for several reasons: They provide benefits that are easily communicated, typically provide no investment risk to employees, and maintain the funding flexibility inherent in a defined benefit plan. However, they also present several tradeoffs: Smaller Accruals for Older Workers As mentioned earlier, unless some type of transition benefits are provided, older employees are likely to receive smaller accruals for their remaining years, regardless of whether a “wear-away” provision (described below) exists. Preretirement Income Replacement Although their understanding of current pension wealth and future increments will no doubt improve vis-à-vis the previous final-average plan, employees actually may be more uncertain about how their future

Employee Benefit Research Institute

benefits will relate to their future earnings after conversion to a cash balance plan. For example, a final-average plan that pays 2 percent of an employee’s average earnings during his or her last three years of service, by definition, replaces 50 percent of preretirement earnings after 25 years of service.21 However, to understand the extent to which cash balance benefits will replace preretirement earnings is far more difficult, since cash balance plans are a type of a career-average formula that provides interest credits that are likely tied to some external financial market vehicle and/or index. Lump-Sum Distributions Due to the increased likelihood that participants in a cash balance plan will end up with a LSD as opposed to a lifetime annuity, it is more likely that they will face a longevity risk (outliving their retirement assets) in addition to a post-retirement investment risk (not earning enough on their investments after leaving work to maintain their retirement assets). It should be noted, however, that with some exceptions, cash balance plans are required to offer annuities as an option to their participants, and it appears that there is an increasing propensity for traditional finalaverage defined benefit plans to offer LSDs and for participants to choose them when offered (Watson Wyatt, 1998). Also, even though cash balance plans communicate benefits in terms of a lump-sum account balance, at least some of them limit the ability of employees to cash out their accounts in a LSD.

Key Issues In recent years, there has been a flurry of press accounts, court cases, and legal and regulatory activities with respect to cash balance plans, specifically as they relate to conversions from existing final-average plans. This section provides some insight into each of these in an attempt to clarify some of their more complex and controversial concepts. Do Cash Balance Plans Result in Cost Savings to the Sponsor? It is certainly possible for conversion to a cash balance plan to result in lower long-term pension expense, depending on the generosity of the new plan relative to the existing plan. In essence, this is no different than switching from a defined benefit to a defined contribution plan, and similar projections would need to be applied to determine if this were the case (VanDerhei 1985). However, even if such a calculation was performed on two retirement plans, it would not necessarily indicate the

73

74

Retirement Security in the United States

extent of cash balance savings, if any, since any savings due to cash balance plan conversion may be offset by other increases in benefits or compensation.22 Assuming such a calculation was performed, the cash balance plan may also prove to be more expensive than originally calculated if turnover is higher than assumed. This would result from plan assets being reduced below expected levels, and the spread between the accrual in the plan and the actual fund performance may be a factor in increased costs.23 Turnover could increase due to future labor patterns that impact all employers, but it might also increase as a direct consequence of providing a more level benefit accrual over time that decreases the “job lock” attributes of the existing plan. However, there may also be short-term abnormalities in the pension cost and/or expense structure resulting from the conversion. In essence, the claims of cost savings from a conversion to a cash balance plan may be at least partially due to a timing issue under the accounting and/or funding rules required for all defined benefit plans (including cash balance plans). Although the calculations are complex, one of the driving forces behind this short-term cost reduction involves the computation of the cost of accruing a benefit based on career-average pay (the cash balance plan) for one based on final-average pay under the previous plan (Demby, June 1999).24 Transition/Grandfathering Several transition methods are available to a sponsor that chooses to mitigate the financial impact that may result in a switch from a traditional final-average plan to a cash balance plan (Rappaport, Young, Levell, and Blalock, 1997). A PricewaterhouseCoopers survey of about 75 cash balance conversions reveals that in almost all cases the employer provided transition provisions beyond the legally required minimums (Sher, 1999). Wear-away If a final-average plan is converted to a cash balance plan, the initial value of a participant’s cash balance account may be set at less than the value of benefits accrued under the previous plan. However, it is important to note that this may not reduce or take away previously earned benefits. It may mean, though, that initially some workers won’t accrue any new benefits until the pay and interest credits to their hypothetical accounts bring the account balances up to the value of the old protected benefits; it is this waiting period that workers must “wear-away” until the new benefit catches up to the old benefit that gives rise to the term. Employers have flexibility in how they credit workers for the value of

Employee Benefit Research Institute

their benefits, and this result could be obtained by computing the opening balance of a participant’s cash balance plan by using a discount rate that is higher than the current 30-year Treasury bond rate.25

Pension Protection Act of 2006 The Pension Protection Act clarifies three of the major uncertainties for plan sponsors considering a cash balance plan. The law’s provisions are prospective only and thus would provide no clarification of the legal status of hybrid plans for past years. Age discrimination The Act clarifies, on a prospective basis, that hybrid plans are not inherently age-discriminatory, as long as benefits are fully vested after three years of service and interest credits do not exceed a market rate of return. Conversion requirements The Act would prohibit, on a prospective basis, any wearaway upon the conversion of a defined benefit plan to a hybrid plan. Thus, an employee’s benefit under the hybrid plan could not be less than the sum of (i) the employee’s benefit under the prior-plan formula as of the date of conversion, plus (ii) the employee’s hybrid plan benefit determined on the basis of service after the conversion. The plan would not have to adjust the prior-plan benefit to reflect future pay growth. The wearaway prohibition would apply to both normal and early retirement benefits. Whipsaw relief When a participant in a hybrid plan elects a lump-sum distribution, the plan may distribute just the participant’s hypothetical account balance, as long as the plan’s interest credit does not exceed a market rate of return. This provision would apply to distributions made after the date of enactment.

75

V

Big Picture Overview of How Retirement “Security” Will be Delivered in the Future

77

Potential Modifications to Social Security, Medicare and Medicaid

T

he federal and state governments are faced with financial challenges in funding the benefits currently scheduled for the elderly. The Social Security program (which provides benefits for the elderly and disabled) is projected to have a shortfall in revenues to pay current law benefits. Medicare (the federal health care insurance program for the elderly and disabled) also is projected to have insufficient revenues to pay for current-law benefits, while state budgets are being strained by the continued growth in Medicaid (the federal-state health care program for poor). Social Security In 1990, the projected 75-year actuarial balance for the Old-Age, Survivors, and Disability Insurance (OASDI) program was –0.91 percent of taxable OASDI payroll, meaning that the OASDI or Social Security tax would need to be increased by 0.91 percentage points to achieve actuarial balance over the next 75 years from 1990 (Figure 34). This 75-year deficit grew to more than 2 percent of taxable payroll before improving to just under 2 percent by 2005. However, this would still be a significant increase of the current taxes—or a corresponding cut in benefits—in order to cover this deficit. On top of the projected deficit, the amount of preretirement income a worker can expect at retirement from Social Security at specific ages of commencing benefits is going to decline due to the scheduled increases in the normal retirement age. For those age 65 in 2005 and starting to receiving benefits in that year and having made the equivalent of the average wage index over the course of their working career, the replacement rate of income just prior to turning age 65 from Social Security is 42.2 percent (Figure 35). For those turning age 65 in 2025 with the same earnings, the expected replacement rate is 36.3 percent. The replacement rates from Social Security are lower for higher earners and higher for lower earners.

Big Picture Overview of How Retirement “Security” Will be Delivered in the Future

79

Percent of OASDI Taxable Payroll

1990

-0.91%

1991

-1.08%

1992

-1.46%

1993

-1.46%

1994

-2.13%

1995

-2.17%

1996

-2.19%

1997

-2.23%

1998

-2.19%

1999

-2.07%

2000

-1.89%

2001

-1.86%

2002

-1.87%

2003

-1.92%

2004

-1.89%

Source: 1990-–05 Annual Reports of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance (OASDI) Trust Funds, available at www.ssa.gov/OACT/TR/index.html.

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

Figure 34 Actuarial Balance of the Federal Old-Age and Survivors Insurance and Disability Insurance (OASDI) Trust Funds, 1990–2005

2005

-1.92%

80 Retirement Security in the United States

Percent of Earnings Just Before Retirement

20%

25%

30%

35%

40%

45%

50%

55%

60%

2025

24.0%

30.1%

36.3%

49.0%

Source: 2005 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance (OASDI) Trust Funds available at www.ssa.gov/OACT/TR/TR05/index.html.

Maximum Earner (Equal to maximum OASDI contribution and benefit base)

2020

25.7%

32.1%

38.7%

52.2%

Medium Earner (100 percent of average wage index)

2015

26.0%

32.3%

38.9%

52.5%

High Earner (160 percent of average wage index)

2010

26.3%

32.0%

38.6%

52.1%

Low Earner (45 percent of average wage index)

2005

29.7%

35.3%

42.2%

56.9%

Figure 35 Replacement Rate of the Social Security Retirement Benefit for Those Retiring at Age 65 From 2005–2025, by Various Steady Earnings Histories

Employee Benefit Research Institute 81

82

Retirement Security in the United States

Medicare and Medicaid After the decline in Medicare expenditures in 1998, the growth in these expenditures increased to 10.6 percent in 2001 (Figure 36). The growth rate moderated modestly to the 7-percent to the 8-percent range through 2004. Rates are projected to be in that range through 2014 with the exception of 2006, when Medicare expenditures are projected to increase about 26 percent due to the addition of the prescription drug program in that year. For each year since 1999, Medicaid expenditures have grown by at least 8 percent. After the expenses of dual-eligible Medicare/Medicaid recipients are moved to Medicare in 2006, Medicaid expenditures are projected to grow in the 8-percent to 9-percent range annually through 2014. Medicaid expenditures grew faster in each year from 1998 to 2003, before Medicare’s growth rate surpassed Medicaid’s rate in 2004. Given these projected growth rates in expenditures and the increasing number of Americans reaching age 65 who are expected to live longer, the Medicare Hospital Insurance Trust Fund is estimated to be facing an actuarial balance of –3.09 percent of taxable payroll, which is a significant increase from the projected balance in 2002 of –2.02 percent (Figure 37). This means that the payroll tax for Medicare would need to be increased by 3.09 percentage points—or benefits would have to be cut by an equivalent amount—in order for the revenues to match expenses for the program over the next 75 years. This level of tax increase would more than double the current payroll tax rate for Medicare. The projected growth in these expenditures could present serious issues for the federal and state governments. According to a Congressional Budget Office (CBO) study, depending on the spending path of expenditures for Medicaid and Medicare, these expenditures could amount to from 5.3 percent of gross domestic product (GDP) to 21.9 percent of GDP by 2050 (Figure 38). For comparison, Social Security expenditures are projected to account for between 6.3 percent to 6.6 percent of GDP by 2050 under the same sending path assumptions. Consequently, given the funding challenges facing government programs and employers, individuals are also likely to be faced with increased challenges of funding a consistent level of expenditures in retirement. Therefore, significantly more preparation and thought must occur at each level in order to maintain a standard of living of all retired Americans.

1998

1999

2000

2001

Actual

2002

2003

2004

2005

Year

2006

2007

2008

2009

Projections

2010

Medicaid

Medicare

2011

2012

2013

2014

Source: Centers for Medicare & Medicaid Services, Office of the Actuary, http://www.cms.hhs.gov/NationalHealthExpendData/02_NationalHealthAccountsHistorical.asp#TopOfPage and http://www.cms.hhs.gov/NationalHealthExpendData/03_NationalHealthAccountsProjected.asp#TopOfPage Note: The 2006 projections include the addition of the Medicare Part D drug program into Medicare, which includes shifting some drug expenditures from Medicaid to Medicare.

-5%

0%

5%

10%

15%

20%

25%

30%

Figure 36 Actual and Projected Percentage Increases in Medicare and Medicaid Personal Health Care Expenditures, 1998–2014

Employee Benefit Research Institute 83

Percent of Taxable Payroll

-3.5%

-3.0%

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

2003

-2.40%

Year

2004

-3.12%

2005

-3.09%

Source: 2002-2005 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, www.cms.hhs.gov/ReportsTrustFunds/01_Overview.asp

2002

-2.02%

Figure 37 Medicare Hospital insurance (HI) Trust Fund Actuarial Balance, 2002–2005

84 Retirement Security in the United States

Percentage of GDP

0%

5%

10%

15%

20%

25%

7.0%

6.3%

Social Security

4.2%

5.9% 5.0%

9.2%

2050

12.6%

6.0%

6.4%

Social Security

4.2%

Expenditure and Spending Path

Intermediate

Medicare and Medicaid

2030

High

Medicare and Medicaid

5.3%

12.0%

6.6%

Social Security

4.2%

6.0%

Source: Congressional Budget Office, "The Long-Term Budget Outlook." A CBO Study , December 2005 http://www.cbo.gov/showdoc.cfm?index=6982&sequence=0

Low

Medicare and Medicaid

4.7%

6.2%

2010

21.9%

Figure 38 Projected Percentage of Gross Domestic Product (GDP) of Medicare and Medicaid and Social Security Expenditures, by Spending Path Assumption, 2010, 2030, and 2050

Employee Benefit Research Institute 85

86

Retirement Security in the United States

The Impact of Moving to a Reduced Social Security Benefits Scenario The impact of moving to a reduced Social Security benefits scenario, assuming housing equity is never liquidated, can be seen by comparing Figure 39 with Figure 14. Given their increased reliance on Social Security benefits as a portion of their overall retirement income, one would expect to observe a larger increase in the median compensation percentages for single females than for single males and for those in the lower income quartiles, cet. par. Comparing the unconstrained groupings in the 1941-1945 birth cohort shows this is exactly what is occurring in the model. For example, the median for single females versus the single males in the third and fourth income quartiles for this birth cohort shows the following increases (in percentages of added compensation that must be saved annually until retirement for a 75 percent chance of covering basic retirement expenses) as a result of moving to the reduced Social Security benefits: Income Quartile

Single Female

Single Male

3

15.9%-15.5% = 0.4%

11.5%-11.2% = 0.3%

4

6.8%-6.5% = 0.3%

5.4%-5.2%=0.2%

Theoretically, there is an indeterminate impact of birth cohort upon the increase in compensation percentages required to offset the reduced Social Security benefits. Even though the younger birth cohorts experience a larger reduction in benefits, this will be offset by the fact that they have a longer period of time for their savings to accumulate.

The Impact of Assuming Lump-Sum Distributions Are Offered to All Defined Benefit Plan Participants at Retirement and They Are Always Chosen Another plan design trend that has been distressing to those who believe defined benefit plans should provide retirement income for the lifetime of former workers26 is the continuing trend of providing employees with the option of taking a lump-sum distribution at the time of retirement.27 Although this is not expected to have any impact on the amount of retirement income and/or wealth generated during the accumulation phase, there is a possibility that the amount of deficits experienced in retirement by an individual or family will increase under this

1

3

1936–1940

2

4

1

3

1941–1945

2

Source: EBRI-ERF Retirement Security Projection Model.

0%

5%

10%

15%

20%

a 25%

4

1

3

1

3

1951–1955

2

single male

Birth Cohort/Income Quartiles

4

single female

1946–1950

2

family

4

1

3

1956–1960

2

Figure 39 Percentage of Added Compensation That Must Be Saved Annually Until Retirement For a 75% Chance of Covering Basic Retirement Expenses (assumes reduced benefits Social Security and housing equity is never liquidated)

4

a

3

1961–1965

2

4

25% = 25% or more.

1

Employee Benefit Research Institute 87

88

Retirement Security in the United States

arrangement. In an attempt to provide a first-order approximation of the impact this change may have on the ability of an individual or family to accumulate sufficient resources to provide for a 75-percent chance of covering simulated retirement expenses, Figure 40 shows the increase in the median percentage of additional compensation that must be saved annually until retirement (cf with Figure 14). Although there appear to be no well-defined trends with respect to age, income, gender, or family status, the vast majority of cohorts would need to increase their savings rate between 10 percent and 20 percent of the rate needed if all defined benefit participants (other than those in cash balance plans) were assumed to receive an annuity.

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

3

1936–1940

2

4

1

3

1941–1945

2

4

1

3

4

1

3

1951–1955

2

4

1

2

3

1956–1960

Birth cohort/Income quartile

1946–1950

2

Source: EBRI-ERF Retirement Security Projection Model.®

1

4

1

3

1961–1965

2

a

4

3

1966–1970

2

4

Single male

Family

Single female

Lump-sum distributions.

1

Figure 40 Increase in Median Percentage of Additional Compensation That Must Be Saved Annually Until Retirement for a 75% Chance of Covering Simulated Expenses, as a Result of Assuming a All Defined Benefit Participants Take LSDs at Retirement

Employee Benefit Research Institute 89

VI

Are Employees Overwhelmed With Choices?

91

A Wide Variety of Savings Tools for Health Care and Retirement Retirement

D

efined Benefit Plans—Although private defined benefit plans do not typically involve employee choice,28 the presence of such a retirement vehicle undoubtedly influences the employee’s decisions with respect to participation and contribution decisions (and perhaps asset allocation decisions) with respect to individual account plans. The types of defined benefit plans provided by private employers will typically fall into one of the following four categories: 

A plan that bases retirement benefits on final average salary.



A plan that bases retirement benefits on career average salary.



A plan that provides a flat dollar amount per year of participation.



A hybrid plan such as a cash balance plan (see Chapter 4).

In addition to these various types of benefit formulae, defined benefit plans are often integrated with Social Security benefits either through an explicit offset for the employee’s monthly Social Security benefit or though a formula that provides larger accruals (or equivalently, less of an offset) for average compensation greater than a specified threshold. 401(k) Plans—With the introduction of Sec. 401(k) to the Internal Revenue Code in 1978 and the issuance of the proposed regulations in November 1981, many private employers were in a position to offer their employees the opportunity to defer salary and/or profit-sharing allocations into a 401(k) plan without current federal income taxation on those amounts. The amount of revenue loss from these arrangements increased rapidly in the early part of the 1980s, and the federal government responded by adopting several constraints that may limit the amount that an employee may contribute:29 Are Employees Overwhelmed With Choices?

93

94

Retirement Security in the United States



Sec. 402(g) limits the employee’s before-tax contribution to $15,000 in 2006 (this number is scheduled to be indexed for inflation in future years).30



If the employee is considered to be highly compensated31 he or she may face additional plan-specific constraints known as an ADP test if the non-highly compensated employees of the plan do not have sufficiently high participation and/or contribution rates.

While employees have often had the opportunity to contribute after-tax amounts beyond the 402(g) limits (and sometimes beyond that allowed by the ADP test),32 beginning in 2006, employers may allow their employees to have a future choice by introducing a so-called Roth 401(k) option to the plan. Although amounts contributed in this fashion are after-tax by definition, they provide the potential advantage that the entire benefit amount (principal plus investment income) may be distributed tax-free if certain requirements are satisfied. 

403(b) and 457 Plans—Certain employers (typically nonprofit entities) are allowed to have Internal Revenue Code (IRC) 403(b) Tax-Sheltered Annuity plans: A public school, college or university.



A church.



A public hospital.



A charitable entity tax-exempt under IRC Sec. 501(c)(3).

Basically, 403(b) plans are similar to 401(k) plans maintained by for-profit entities. Just as with a 401(k) plan, a 403(b) plan lets employees defer some of their salary. In this case, their deferred money goes to a 403(b) plan sponsored by the employer. This deferred money generally is not taxed by the federal government until distributed. Plans of deferred compensation described in IRC Sec. 457 are available for certain state and local governments and non-governmental entities tax exempt under IRC Sec. 501. They can be either eligible plans under IRC Sec. 457(b) or ineligible plans under IRC Sec. 457(f ). Plans eligible under 457(b) allow employees of sponsoring organizations to defer income taxation on retirement savings into future years. Ineligible plans may trigger different tax treatment under 457(f ).

Employee Benefit Research Institute

Individual Retirement Accounts (IRAs)—An IRA is a personal savings plan that provides an individual with tax advantages for setting aside money for retirement. Contributions to an IRA may be fully or partially deductible, depending on which type of IRA is chosen and the individual and/or family circumstances. Similar to 401(k) plans, the amounts in an IRA (including earnings and gains) are generally not taxed until distributed. If the Roth version of an IRA is used, amounts are not taxed at all if distributed according to the rules. If contributions are made exclusively to traditional IRAs, the limit for 2006 is $4,000 ($5,000 for those ages 50 or older). If contributions are made exclusively to Roth IRAs, the same limits may apply; however, if the modified AGI is above a certain amount, the contribution limit may be reduced. If the individual is covered by a retirement plan at work, his or her deduction for contributions to a traditional IRA will be reduced or eliminated if their modified adjusted gross income (AGI) is in excess of specified amount. Health33 Health Savings Accounts—A health savings account (HSA) is a taxexempt trust or custodial account that an individual can use to pay for health care expenses. Contributions to the account are deductible from taxable income, even for individuals who do not itemize, and distributions for qualified medical expenses are not counted in taxable income. Tax-free distributions are also allowed for certain premiums (this is discussed in more detail below). HSAs are owned by the individual with a high-deductible health plan and are completely portable. There is no use-it-or-lose-it rule associated with HSAs, as any money left in the account at the end of the year automatically rolls over and is subsequently available. A bank, insurance company, or other nonbank trustee approved by the Internal Revenue Service (IRS) must administer the HSA. High-Deductible Health Plans (HDHPs)—In order for an individual to qualify for tax-free contributions to an HSA, the individual must be covered by a health plan that has an annual deductible of not less than $1,000 for self-only coverage and $2,000 for family coverage. Certain preventive services can be covered in full and are not subject to the deductible. The out-of-pocket maximum may not exceed $5,000 for self-only coverage and $10,000 for family coverage, with the deductible counting toward this limit. The minimum allowable deductible and

95

96

Retirement Security in the United States

maximum out-of-pocket limit will be indexed to inflation in the future. Network plans may impose higher deductibles and out-of-pocket limits for out-of-network services. Contributions to an HSA Both individuals and employers are allowed to contribute to an HSA. Contributions are excluded from workers’ taxable income if made by the employer, and deductible from adjusted gross income if made by the individual. The maximum annual contribution is $2,700 for self-only coverage and $5,450 for family coverage in 2006. But the maximum permissible contribution cannot exceed the plan deductible. This means that an individual with a $1,000 deductible is not allowed to contribute more than $1,000 to an HSA. Future contribution limits will be indexed to inflation.34 To be eligible for an HSA, individuals may not be enrolled in other health coverage, such as a spouse’s plan, unless that plan is also a highdeductible health plan. However, individuals are allowed to have supplemental coverage without a high-deductible for such things as vision care, dental care, specific diseases, and insurance that pays a fixed amount per day (or other period) for hospitalization.35 Beneficiaries enrolled in Medicare are not eligible to make HSA contributions, although they are able to withdraw money from the HSA for qualified medical expenses and certain premiums.36 Individuals also may not make an HSA contribution if claimed as dependents on another person’s tax return. Individuals who have reached age 55 and are not yet enrolled in Medicare may make catch-up contributions. In 2006, a $700 catch-up contribution is allowed. A $1,000 catch-up contribution will be phased in by 2009.37 Distributions From an HSA Distributions from an HSA can be made at any time. An individual need not be covered by a high-deductible health plan to withdraw money from his HSA (although he must have been covered by a high-deductible health plan at the time the funds were placed in the HSA). Distributions are excluded from taxable income if they are used to pay for qualified medical expenses as defined under IRC Sec. 213(d). Distributions for premiums for COBRA (Consolidated Omnibus Budget Reconciliation Act of 1985), long-term care insurance, health insurance while receiving unemployment compensation, and insurance while eligible for Medicare other than for Medigap, are also tax-free. This means that distributions used to pay Medicare Part A or B, MedicareAdvantage plan premiums,

Employee Benefit Research Institute

and the employee share of the premium for employment-based retiree health benefits are allowed on a tax-free basis. Distributions for nonqualified expenses are subject to regular income tax as well as a 10 percent penalty, which is waived if the owner of the HSA dies, becomes disabled, or is eligible for Medicare. Individuals are able to roll over funds from one HSA into another HSA without subjecting the distribution to income and penalty taxes as long as the rollover does not exceed 60 days. Rollover contributions from Archer medical savings accounts (MSAs) are also permitted. Earnings on contributions are also not subject to income taxes. Flexible Spending Accounts Flexible spending accounts (FSAs) are a type of benefits cafeteria plan, authorized under IRC Sec. 125 as part of the Revenue Act of 1978. FSAs can be offered on a stand-alone basis or as part of a larger cafeteria plan, under which participants can choose among two or more benefits and cash. FSAs do not need to be paired with a high-deductible health plan. Individuals are eligible for an FSA only if an employer offers it as an option. Contributions to an FSA FSAs are funded through employee pre-tax contributions. Employees must designate their contribution in the year prior to the plan year. Once made, changes are allowed only for certain circumstances, such as a change in family status, plan cost changes, and plan coverage changes. Contributions to FSAs are withheld in equal amounts from each paycheck throughout the plan year, but employers must make the full amount available to the employee at the beginning of the plan year. For example, an employee who chooses to contribute $1,200 to an account will have $100 deducted from his or her paycheck each month, but will have access to the full $1,200 at the beginning of the plan year. If an employee is reimbursed more than he or she has contributed to the account, and then leaves the job, the employer may lose money on the arrangement. This rule is a disincentive for a small employer to offer such an account. While there is no statutory limit on annual contributions to a medical FSA, employers are allowed to set an upper limit, and usually do so to mitigate losses related to turnover. Contributing to an FSA not only reduces salary for federal income tax purposes, but also reduces the wages on which Social Security and Medicare taxes are paid. As a result, employees with earnings below the

97

98

Retirement Security in the United States

Social Security wage base will also pay less in Social Security taxes after the deduction is made for FSA contributions. Employees at all income levels will also pay less in Medicare taxes. The employer’s share of Social Security and Medicare taxes will also be reduced, and this reduction may be large enough to offset the cost of administering the benefit. Distributions From an FSA Distributions from an FSA can be made at any time. Distributions are excluded from taxable income if they are used to pay for qualified medical expenses as defined under IRC Sec. 213(d). Employees forfeit any money left over in the FSA at the end of the plan year; this is known as the “use-it-or-lose-it” rule. Employers can keep the forfeited funds and use them for any purpose, except that the funds cannot be returned to employees who have forfeited them. Medical Savings Accounts A medical savings account (MSA) is a tax-exempt trust or custodial account that an individual can use to pay for health care expenses. Employees are eligible to set up an MSA if employed at a firm with 50 or fewer employees. The self-employed are also eligible. Both must be covered by a high-deductible health plan. High-Deductible Health Plan In order for an individual to qualify for tax-free contributions to an MSA, he or she must be covered by a health plan that has an annual deductible of between $1,700 and $2,600 for self-only coverage and between $3,450 and $5,150 for family coverage. Certain preventive services can be covered in full and are not subject to the deductible. The out-of-pocket maximum may not exceed $3,450 for self-only coverage and $6,300 for family coverage. The allowable deductible range and maximum out-of-pocket limit are indexed to inflation. Contributions to an MSA Both employees and employers are allowed to contribute to a worker’s MSA, but both may not make contributions in the same year. Contributions are excluded from taxable income if made by the employer, and deductible from adjusted gross income if made by the individual. Contributions cannot exceed annual earned income or net self-employment income.

Employee Benefit Research Institute

Distributions From an MSA Distributions are excluded from taxable income if they are used to pay for qualified medical expenses as defined under IRC Sec. 213(d). Distributions for premiums for COBRA, long-term care insurance, and health insurance while receiving unemployment compensation are also tax-free. Distributions for nonqualified medical expenses are subject to regular income tax as well as a 15 percent penalty, which is waived if the owner of the MSA is age 65 or older, becomes disabled, or dies.

Options for Consolidation and Simplification Choice complexity occurs when taxpayers are given numerous subsidies but may only use one or a few of them. This type of complexity has grown into the large number of savings tools for health care and retirement described above. Although more choices may be advantageous to some individuals for tax planning and/or maximum contributions, this increased complexity will certainly have attendant costs with respect to learning about the various requirements for each. William Gale38 notes that in designing tax policy it is not necessarily the case that more options are always worth the added costs. First, the differences in benefits to a household between choosing one of a set of options versus another in the same set may be smaller than the costs of determining which the best option. But of course the household does not know that until it has undertaken the cost. Second, having more choices, for example, with respect to retirement saving, requires more record-keeping by the taxpayer and the government. Although there have been several proposals to modify the complexity of this system, the most recent concerted effort to analyze the costs and benefits of various modifications took place in January 2005 when President George W. Bush created the President’s Advisory Panel on Federal Tax Reform. Since then, the panel has analyzed the current federal income tax system and considered a number of proposals to reform it. One of the recommendations was to combine the different tax provisions for at-work saving, health saving, education saving, and retirement saving into three simple saving plans. Although no changes were suggested for defined benefit plans, defined contributions plans would be consolidated into Save at Work plans that have simple rules and use current-law 401(k) contribution limits.39 IRAs, Roth IRAs, spousal IRAs would be replaced with Save for Retirement accounts ($10,000 annual limit) available to

99

100

Retirement Security in the United States

all taxpayers. Education savings plans and health savings plans would be replaced with Save for Family accounts ($10,000 annual limit); would cover education, medical, new home costs, and retirement saving needs; and would be available to all taxpayers.

VII

An Example of How Employers Can Improve Retirement Security for Their Employees

Automatic Enrollment for 401(k) Participants40

101

A

s the evolution from defined benefit to defined contribution plans (especially those of the salary deferral or 401(k) type) has continued over the last two decades, many policy makers have been concerned that employees may not be taking advantage of the employer sponsored retirement plans soon enough (or in some cases at all) to end up with an adequate retirement income. This appears to be most obvious when looking at the participation rates among young and low income workers. A relative new procedure known as automatic enrollment or negative election has been adopted by some 401(k) sponsors in an attempt to use inertia to the employee’s advantage by initially enrolling them in the 401(k) plan and then giving them the option to opt out at their discretion. Under an automatic enrollment provision, the initial contribution rate is set by the plan sponsor as well as a default investment option. Profit Sharing/401(k) Council of America (PSCA; 2004) reports that 58 percent of plans with automatic enrollment set the employee contribution rate at 3 percent of salary and another 22 percent of plans with automatic enrollment chose 2 percent of salary.41 While this is greater than the zero percent contribution rate that nonparticipants choose, these automatic contribution rates generally are lower than average contribution rates of 401(k) participants in the EBRI/ICI database. The average before tax contribution rate for all groups was in excess of 6 percent with the exception of employees in their twenties earning less than $40,000. In addition, in many cases the default contribution rate in the automatic enrollment plans is lower than the contribution rate needed to take full advantage of the employer match.42

Legislative/Regulatory Environment for Auto-Enrollment In 1998, the IRS issued a ruling clarifying that automatic enrollment in §401(k) plans is permissible for newly hired employees (Revenue Ruling 98-30). The IRS issued a second ruling in 2000 stating that automatic enrollment also is permissible for current employees who have not already enrolled in the plan (Revenue Ruling 2000-8). In 2004, the IRS

An Example of How Employers Can Improve Retirement Security for Their Imployees

103

104

Retirement Security in the United States

published a general information letter in response to a public inquiry that clarified two previously ambiguous points. The letter stated that (1) the amount deducted from the employee’s pay and contributed to the plan can be any amount that is permissible under the plan up to the annual contribution limits under IRC §402(g), and (2) the plan can automatically increase the employee’s contribution over time, such as after each pay raise.43

Cost/Benefit Analysis for Employers Considering AutoEnrollment The potential benefits of adopting automatic enrollment are described in the next section; however, in 2005, only 19 percent of large employers automatically enrolled workers, and the average rate was only 3 percent of annual pay, according to Hewitt Associates.44 While the number of 401(k) sponsors adopting this procedure is certainly increasing (in 1999 only 7 percent of the companies in the Hewitt survey had adopted this procedure), there are several reasons why the majority of plans still do not offer this feature: 

Cost implications.



State labor laws.



Fiduciary exposure.

The potential increase in employer cost will obviously vary with plan design features and the employees’ response to them. Many eligible employees choose not to contribute anything at all and even those who do contribute often fail to take maximum advantage of the matching feature. Increases in participation rates and possibly contribution rates will translate into increased employer contributions to a 401(k) plan with a matching formula. In approximately 30 states, some employers see their state labor laws as potentially restricting their ability to adopt automatic enrollment. The laws require affirmative consent before any amounts can be deducted from an individual’s payroll. In the private sector, some corporations have implemented automatic enrollment based on readings that certain provisions in ERISA override these state laws.45 Some companies are reluctant to either use automatic enrollment or, if they do choose to do so, only provide an extremely conservative investment option as a default for fear that employees who lost money

Employee Benefit Research Institute

would sue. This potential limitation may be mitigated in the near future, though, if the Department of Labor proposes a regulation clarifying that balanced investment funds may be acceptable from a fiduciary standpoint.46 The exemption from fiduciary responsibility would likely not be total: Plan fiduciaries would retain appropriate responsibility for avoiding conflicts of interest, excessive fees, lack of diversification, and imprudent investment choices.47

Simulation Results This section reviews the results of an EBRI/ICI simulation conducted to assess the potential value to employees of 401(k) plan sponsors adopting automatic enrollment. Significant increases in average account balances were found for low-income employees regardless of the choice of the default contribution rate and asset allocation. As the default contribution rate is increased from 3 to 6 percent of compensation and as the default asset allocation is changed from a very conservative selection to one that contains an age-appropriate mix of equities and bonds, the average account balances for all income categories increase. The EBRI/ICI 401(k) Accumulation Projection Model was originally designed to simulate future retirement income at age 65 for 401(k) plan participants who had account balances at year-end 2000. In order that the impact of increasing participation under an automatic enrollment provision could be assessed, we needed to construct a cohort of synthetic non-participants to add to the model. Run under the baseline assumptions, the median replacement rate for the lowest income quartile is 23 percent of preretirement income and 56 percent for the highestincome quartile. The EBRI/ICI model must also know how to estimate which employees will remain with the default choices and for how long. For this set of simulations, the model’s automatic enrollment assumptions are based on an analysis by Choi, Laibson, Madrian, and Metrick (2001 and 2004)48 of data for a company with a default contribution rate initially set at 3 percent of salary and an initial default asset allocation in a money market fund. The employer match rate for the plan was 50 percent of up to 6 percent of pay after one year of employment. Choi et al. (2001) find a positive influence on participation among new employees: initially only about a third of them were participating in the 401(k) plan, while after automatic enrollment this figure increased to 87 percent. Although the impact on participation rates was very positive,

105

106

Retirement Security in the United States

there was an offsetting impact on contribution rates (for those that participated). The most common contribution rate prior to automatic enrollment was 6 percent of compensation (the maximum amount matched by the employer). However, after the automatic enrollment program was implemented, 72 percent of new employees contributed at the default contribution rate (3 percent). There was also a shift towards less aggressive investments after the introduction of automatic enrollment: only 18 percent of participants of new employees had all of their 401(k) balances in the money market fund but this figure increased to 71 percent for those hired after automatic enrollment was installed. In order to forecast the impact of automatic enrollment on a broader population of workers over an entire career, the EBRI/ICI projection model assumes that 401(k) plan sponsors immediately implement the automatic enrollment behaviors described above and compares the replacement rates of four different automatic enrollment scenarios with the replacement rates among all eligible workers without automatic enrollment. The model analyzes two different default contribution rates—3 percent of salary and 6 percent of salary—and two different default asset allocations—a money market fund and a life-cycle fund. For employees assumed to work for 401(k) plan sponsors their entire careers, the impact of automatic enrollment with a 3 percent of salary contribution rate and a money market fund varies from an increase from 23 percent to 37 percent in the median replacement rate for those in the lowest income quartile at age 65 to a decrease from 56 percent to 52 percent for the median replacement rate for those in the highest income quartile. There is a positive impact on the replacement rates for those with lower incomes because the effect of increasing participation rates for these employees more than offsets the potential downside of reducing contribution rates and/or investing more conservatively. In contrast, higher paid employees already have such high participation rates even in the absence of automatic enrollment that there is very little to be gained by increasing participation. Rather, projected replacement rates are reduced by automatic enrollment’s lower contribution rates and/or more conservative investment strategies. When a default contribution rate of 3 percent is paired with a default investment allocation of a life cycle fund, the median replacement rate for the lowest income quartile is estimated to be 42 percent and the figure for the highest income quartile is 57 percent (slightly more than the baseline before automatic enrollment was introduced). When the default con-

Employee Benefit Research Institute

tribution rate is increased to 6 percent of compensation and the default investment is a money market fund the estimated median replacement rates are 45 percent for the lowest income quartile and 58 percent for the highest paid quartile. When both the 6 percent default contribution and the life cycle default investment are assumed, the median replacement rate for the lowest income quartile is estimated to be 52 percent: a 126 percent increase from the rate estimated before the introduction of automatic enrollment! The highest income quartile is estimated to benefit with a 13 percent increase with a median replacement rate of 63 percent.

107

VIII

Defined Benefit Plan Freezes

109

T

he dawn of the new year in 2006 began with an increase of press coverage of a decision among plan sponsors that had actually been quite prevalent in recent years: freezing defined benefit plan accruals for new and/or existing employees. This chapter provides a detailed analysis of how such activity is likely to impact existing employees as a function of plan type and employee demographics.

Financial Consequences of a Pension Freeze This process can be illustrated through the use of a stylized example. Assume an employee begins participating in a defined benefit pension plan at age 30 and plans to retire at age 65. The defined benefit plan is a final average plan that provides a benefit accrual of 1 percent per year of participation times the average of the final three years of compensation. In 2006 when the employee is age 50 and earning $70,000, the defined benefit plan is frozen and replaced with a 401(k) plan. How much will the employee need to accumulate in the plan to end up with the same expected retirement income as he would have had if the plan had not been frozen? If we assume the employee would have a constant salary growth of 3 percent per year, the final year’s salary at age 64 would be equal to $105,881 and the final three-year average compensation would be $102,827. This would have provided a nominal annual annuity of $35,989 beginning at age 65 if the plan had not been frozen. However, the new average compensation is now $67,980 and the accrued benefit is frozen at $13,596 per year beginning at age 65. The difference of $22,393 would need to be made up by purchasing an annuity at age 65 with the proceeds of the 401(k) balance. Assuming that the annuity purchase price at that time is 13.38,49 the employee would have needed to accumulate $299,536 in his 401(k) plan to purchase an annuity to fill in the gap created by the pension freeze. If one assumes the same age-specific asset allocations observed in the EBRI/ICI Participant-Directed Retirement Plan data base from year-end

Defined Benefit Plan Freeze

111

112

Retirement Security in the United States

200450 and a deterministic annual rate of return of 10.4 percent nominal on equities and 5.5 percent nominal on other asset classes, the requisite amount of money would be accumulated during the following 15 years if a total contribution of 12.87 percent of compensation is made at the end of each year to the employee’s 401(k) plan. Whether this money is provided by the employee, the employer, or a combination of the two is irrelevant for purposes of providing the full benefit that was expected prior to the freeze;51 however, in most cases the previous (defined benefit) plan would be noncontributory and a full financial indemnification for the freeze would suggest that the full annual contribution be provided by the employer. While the previous example illustrates the percentage of annual compensation that must be contributed for indemnification of the amount of expected retirement income lost as a result of the pension freeze, the result is extremely sensitive to many of the underlying assumptions that were made. For example, if the assumed rate of return on equities is decreased to 8 percent nominal per year, the percentage of compensation needed to indemnify the individual for the loss in expected retirement income as a result of the pension freeze (hereafter referred to as the “indemnification contribution rate”) would increase from 12.87 percent to 14.3 percent. In a similar manner, if the assumed salary growth rate increased to 4 percent nominal per year, the indemnification contribution rate would increase to 14.6 percent. On the other hand, there are other assumptions that may not be realized in the stylized example that would tend to reduce (rather than increase) the indemnification contribution rate. Perhaps the most obvious example is that very few employees will work for a single employer for their entire careers—which means very few employees will qualify for the full final-average pension benefit potentially attainable.52 If the assumption in the stylized example above is relaxed such that the employee were to change jobs five years after the pension freeze (at age 55), the indemnification contribution rate drops to 7.5 percent per year (this is because the potential final-average pension benefit is substantially smaller due to the job change). Beside the key assumptions of future rates of return on various asset classes and future salary growth, the indemnification contribution rate will depend, among other things, upon: 

The type of defined benefit plan that was previously sponsored. Final-average plans (and to a lesser extent, career-

Employee Benefit Research Institute

average plans) tend to grow faster toward the end of the employee’s active working career than the other plan types and, (all else equal), should have a larger indemnification contribution rate for older and/or longer-tenured employees; the bigger the potential pension benefit, the more would have to be made up through 401(k) contributions. Cash balance pension plans (described earlier in this report) tend to accumulate benefits in a fashion that is similar to 401(k) plans and the indemnification contribution rate will be determined by the relative size of the cash balance pay credits and the spread between the assumed interest credit for the cash balance plans and the assumed rate of return for the 401(k) plans. One particularly vexing problem in estimating the indemnification contribution rate for cash balance plans deals with the choice of an assumed interest credit in the future. Although the distribution of current rates can be sampled, assumptions regarding future rates may be more likely to trend toward rates of return on longterm Treasuries. This problem is addressed later in the report by analyzing cash balance plans with both current and longterm rates.53 

Generosity parameters of the defined benefit plan. Within any plan type, there are a variety of plan design considerations that may make the plan more or less valuable to the employee. For example, in the stylized example above, the 1 percent of final-average, three-year compensation could have been 1.5 percent. Alternatively the averaging period might have been five years or the plan benefits might have been integrated with Social Security. As described above, the higher the relative value of the defined benefit pension plan, the higher the indemnification contribution rate.



Age of initial participation in the current plan, age at time of pension freeze, and age at time of job change following the pension freeze. Mathematically, the indemnification contribution rate will be determined by each of these values and can be analyzed by looking at the impact of each of the following on the indemnification contribution rate:

1. Length of time with the employer after the pension freeze until job change.

113

114

Retirement Security in the United States

2. Remaining years to retirement after leaving the current job. 3. Years already worked on the current job. 4. Age at the time of the pension freeze. With the exception of the first variable above, there should be a straightforward relationship between the values of each of these and the indemnification contribution rate. The second variable should have a negative association with the indemnification contribution rate, given that the longer the time after the job change, the longer the period of time the 401(k) balance can accumulate until retirement age without having a commensurate increase in the defined benefit accrual under the assumption that the plan had not been terminated. The last two variables should have a positive association with the indemnification contribution rate: The first is indeterminate (since there are two opposing forces at work: the longer the employee was going to work after the point of the pension freeze, the larger the future defined benefit accrual; however, the longer period also gives additional time for 401(k) contributions to accumulate under the assumption that the defined benefit pension has been frozen).

Analyzing the Financial Consequences of a Pension Freeze This section analyzes the financial consequences of a pension freeze for the general population of participants in private defined benefit plans in 2006. This is accomplished by utilizing the EBRI-ERF model described in the appendix. Briefly, the model takes the current population of workers in the private sector in 2006, statistically attributes whether or not they are participating in a defined benefit plan and, if so, what type of plan and the attendant generosity parameters. The model incorporates a stochastic job tenure algorithm that provides information on how long the employee has already participated in the defined benefit plan and how much longer after 2006 they will remain with the employer. With this information, the reduction in the future estimated defined benefit income as a result of a pension freeze in 2006 can be estimated, and the indemnification contribution rate for each defined benefit participant can be determined. As shown above, this calculation will be sensitive to the choice of the rates of return on various asset classes—and it is clear that there is no consensus on what future returns in the financial markets will be for the next 30 years. Therefore, the EBRI-ERF model suppresses the stochastic

Employee Benefit Research Institute

rate of return mechanism typically employed by this type of analysis and substitutes a constant rate of return of either 4 percent nominal per year or 8 percent. This allows readers to choose which rate they believe is more likely for the future and use the corresponding set of results. In addition, the EBRI-ERF model makes one more modification before undertaking this analysis. It is a well-known fact that job tenure is larger for defined benefit participants than for either defined contribution participants or workers in general, given the financial consequences of job change upon employees participating in a final-average defined benefit pension plan. Therefore, the typical tenure distributions are replaced with those for participants exclusively in defined contribution plans to account for the increased job mobility that is likely to accompany the pension freeze.

Overall Results by Plan Type The median indemnification contribution rate for a career-average defined benefit pension plan is 11.6 percent, assuming a 4 percent rate of return. An indemnification contribution rate of 18.8 percent would be sufficient to cover 75 percent of the employees covered by this type of plan. The median rate for a final-average plan is larger, as expected: 13.5 percent, and the threshold rate for the 75th percentile increases to 21.0 percent. Cash balance plans have a median indemnification contribution rate of 4.6 percent, with a 75th percentile threshold rate of 6.3 percent using the current interest credits. These values increase to 5.7 percent and 7.3 percent if, instead, the cash balance plans are assumed to credit interest at the intermediate long-term assumption for the interest rate of the Treasury special public-debt obligation bonds issuable to the OASDI trust funds, as specified in the 2005 Trustees of the OASDI Trust Funds Report (5.8 percent). If the rate of return assumption is increased to 8 percent nominal, the median indemnification contribution rate for a career-average defined benefit plans is 6.6 percent. An indemnification contribution rate of 14.8 percent would be sufficient to cover 75 percent of the employees covered by this type of plan. The median for a final-average plan is 8.1 percent and the 75th percentile threshold increases to 16.0 percent. Cash balance plans have a median indemnification contribution rate of 2.7 percent, with a 75th percentile threshold of 4.5 percent using the current interest credits. These values increase to 3.1 percent and 5.2 percent if the cash balance plans are assumed to credit interest at 5.8 percent.

115

116

Retirement Security in the United States

The cumulative distribution functions of the indemnification contribution rate needed to offset the impact of a pension freeze in 2006, by pension plan type, is shown in Figures 41 and 42 for rates of return of 4 and 8 percent respectively.

Results by Age and Plan Type Although the median indemnification contribution rate for careerand final-average plans at a 4 percent rate of return were 11.6 percent and 13.5 percent respectively, these values ignore the impact of the employee’s age on the rates. Figure 43 shows that the median rates increase substantially with the age of the employee. For career-average plans, the medians increase from 5.1 percent for those currently age 30-34 to 20.1 percent when they are 60-64. Final-average plans have an even larger increase: from 3.9 percent to 22.4 percent. Cash balance plans display little variation by age regardless of the interest credit chosen. Figure 44 shows similar results when an 8 percent rate of return is chosen: the indemnification contribution rates for career-average plans increase from 1.4 percent for the youngest workers analyzed to 16.6 percent for the oldest cohort. The rates for final-average plans increase from 1.1 percent to 18.6 percent for the respective age cohorts, while cash balance plans remain much flatter than the other two types, but no longer appear age-invariant: For cash balance plans, the rates increase from 1.1 percent to 4.8 percent for the young/old age cohorts when current interest credits are used, and 1.7 percent to 5.4 percent when long-term rates are assumed.

Results by Tenure With the Current Employer at the Time of the Pension Freeze and by Plan Type The number of years already worked with the current employer at the time of the pension freeze can have two potential influences on the indemnification contribution rate for non-cash balance plans. First, the longer an employee has worked with an employer at the time of the freeze, the shorter the period of time they will have to accumulate balances in the 401(k) plan after the subsequent job change with respect to the current employer. Second, the longer an employee has worked for an employer at the time of the freeze, the more years of accruals will be affected by the reduced salary average in a final-average plan. Careeraverage plans will be affected by the first influence only, while final-average plans will be affected by both.

Cumulative Percentage of Employees

0–5%

5–10%

10–15%

15–20%

15–20%

Percentage of Compensation

Source: Author's tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

20–25%

25–30%

30–35%

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Figure 41 Cumulative Distribution Function of the Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type (assumes 4% annual rate of return)

Employee Benefit Research Institute 117

Percentage of Employees

0–5%

5–10%

10–15%

15–20%

20–25%

25–30%

30–35%

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Percentage of Compensation

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Figure 42 Cumulative Distribution Function of the Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type (assumes 8% annual rate of return)

35–40%

118 Retirement Security in the United States

Percentage of Compensation

30–34

35–39

40–44

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

45–49 Current Age

Source: Author's tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

5%

10%

15%

20%

25%

50–54

55–59

Figure 43 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Current Age (assumes 4% annual rate of return)

60–64

Employee Benefit Research Institute 119

Percentage of Compensation

30–34

35–39

40–44

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Current Age

45–49

Source: Author's tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

50–54

55–59

Figure 44 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Current Age (assumes 8% annual rate of return)

60–64

120 Retirement Security in the United States

Employee Benefit Research Institute

Figure 45 shows the impact of this factor at an assumed rate of return of 4 percent. The career-average indemnification contribution rate increases from 10.7 percent for employees with zero to four years of prior tenure with the employer, to 14.0 percent for those with 15 or more years. The final-average indemnification contribution rate for similar groups increases from 10.7 percent to 17.3 percent. Figure 46 shows the same results assuming an 8 percent rate of return. The career-average indemnification contribution rate increases from 5.2 percent for employees with zero to four years of prior tenure with the employer, to 10.8 percent for those with 15 or more years. The finalaverage indemnification contribution rate for similar groups increases from 6.0 percent to 12.9 percent.

Results by Remaining Years to Retirement After Leaving Job With the Current Employer and by Plan Type The remaining years to retirement after leaving the job with the current employer is expected to affect the indemnification contribution rate because this represents the remaining time employees would have to accumulate investment returns in a 401(k) plan, assuming the pension freeze had occurred with a corresponding benefit accrual. In fact, the reason the current-age graphs (above) had the relationship they did was because they were acting as a proxy for this variable. At a 4 percent rate of return (Figure 47), the indemnification contribution rate for a career-average plan decreases from 18.9 percent for employees with zero to four years of remaining years to retirement, to 3.9 percent for those with 30 or more years. The final-average indemnification contribution rate for similar groups increases from 20.7 percent to 3.5 percent. At an 8 percent rate of return (Figure 48), the career-average indemnification contribution rate decreases from 14.9 percent for employees with zero to four years of remaining years to retirement to 1.1 percent for those with 30 or more years. The final-average indemnification contribution rate for similar groups decreases from 15.7 percent to 1.0 percent These figures demonstrate the strong impact that a pension freeze has on older workers (who have little time left in their working careers to accumulate assets in a defined contribution plan), especially compared with younger workers (who have much more time).

121

Percentage of Compensation

0–4

10–14

Years Already Worked on the Job With the Current Employer

5–9

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Source: Author's tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

Figure 45 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Years Already Worked on the Job With the Current Employer (assumes 4% annual rate of return)

15 or More

122 Retirement Security in the United States

Percentage of Compensation

0–4

10–14

Years Already Worked on the Job With the Current Employer

5–9

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

12%

14%

15 or More

Figure 46 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze in 2006, by Pension Plan Type and Years Already Worked on the Job With the Current Employer (assumes 8% annual rate of return)

Employee Benefit Research Institute 123

Percentage of Compensation

0–4

10–14

15–19

20–24

25–29

30 or More

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Remaining Years to Retirement after Leaving Job With the Current Employer

5–9

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

5%

10%

15%

20%

25%

Figure 47 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years to Retirement after Leaving Job With the Current Employer (assumes 4% annual rate of return)

124 Retirement Security in the United States

Percentage of Compensation

0–4

5–9

10–14

15–19

20–24

Remaining Years to Retirement After Leaving Job With the Current Employer

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

25–29

30 or More

Cash Balance–Current Interest Rates Cash Balance–Long-Term Interest Rates

14%

12%

Final Average

Career Average

16%

18%

Figure 48 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years to Retirement after Leaving Job With the Current Employer (assumes 8% annual rate of return)

Employee Benefit Research Institute 125

126

Retirement Security in the United States

Results by Tenure With Employer After the Pension Freeze Until Job Change and by Plan Type The number of years that an employee remains with an employer after the pension plan is frozen will affect the indemnification contribution rate. For final- and career-average plans, the indemnification contribution rate is likely to increase for several years until the point where the 401(k) accumulations can offset the incremental benefit accruals. For example, in Figure 49 the indemnification contribution rate at an 8% rate of return for final-average plans peaks at 13.0 percent for employees with 10-14 remaining years with the employer after the freeze. After that point, additional 401(k) accumulations more than make up for the extra pension accruals that would have been accumulated if the plan was not frozen, so the indemnification contribution rate begins to decrease: 12.3 percent for workers with 15-19 years of tenure and 10.9 percent for those with 20 or more years. A similar situation arises for the career-average plans; however, it does not peak until 15-19 years after the pension freeze. Figure 50 shows the same situation for a 4 percent rate of return; however, with the reduced amount of investment income, the period for the downturn in the indemnification contribution rate curve is delayed. For example, the final-average indemnification contribution rate peaks at 18.3 percent for employees 15-19 years after the pension freeze and decreases nominally to 18.0 percent for those with 20 or more years of tenure. As defined benefit pension sponsors continue to freeze benefit accruals for new and/or current employees and substitute either new or enhanced 401(k) plans, many observers will be concerned whether the total expected retirement income from the combination of the frozen defined benefit and the new/additional 401(k) balances will equal or exceed what the employees might have thought they would receive from the original defined benefit plan, assuming it continued without modifications. As Figures 41 and 42 show, there is tremendous variability regarding what it would take to financially indemnify an employee for such a freeze. Because workers affected by a pension freeze vary greatly by age, salary, and job tenure; by the specific provisions and formula in the types of retirement plans they are covered by (both pension and 401(k)); and by the underlying economic assumptions that are used to estimate the effects of a pension freeze, there is fundamentally no simple answer to the question.

Employee Benefit Research Institute

This analysis has used the EBRI/ERF model (Figures 43-50) to show how results will differ depending on plan type, assumed rate of return, and a variety of tenure factors. This analysis provides a construct for employers to estimate the relative impact of certain demographics on those covered by defined benefit pension plans in general. However, each plan contains design features that make it unique, and individual analysis of costs and benefits needs to be done for each employer contemplating such a move.

127

Percentage of Compensation

0–4

5–9

10–14

15–19

20 or More

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Remaining Years on the Job With the Current Employer

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

12%

14%

Figure 49 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years on the Job With the Current Employer (assumes 8% annual rate of return)

128 Retirement Security in the United States

Percentage of Compensation

0–4

5–9

10–14

15–19

20 or More

Cash Balance–Long-Term Interest Rates

Cash Balance–Current Interest Rates

Final Average

Career Average

Remaining Years on the Job With the Current Employer

Source: Authors' tabulations from the EBRI/ERF Retirement Income Projection Model.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

Figure 50 Median Percentage of a Worker’s Annual Pay Needed to Offset the Impact of a Pension Freeze, by Pension Plan Type and Remaining Years on the Job With the Current Employer (assumes 4% annual rate of return)

Employee Benefit Research Institute 129

IX

Conclusion

131

T

his report documented the long term restructuring of the nation’s retirement income programs, the changing pattern of retiree income across the age spectrum, the amounts that would need to be saved beyond current savings for today’s workers to retire with sufficient funds to meet basic living expenses, and the amounts that would need to be set aside in order for workers who experience a defined benefit plan freeze to have the same level of retirement income that the former plan would have provided. The report also underlines the complexity of these issues and the degree to which the income, age and tenure of workers affects participation in the system, the probabilities of successful outcomes, and the implications of system change for the individual. For retirees the report presents the relative role of Social Security and other income sources in retirement, and presents a picture of the tremendous changes that take place in this mix as retirees age. As the nation begins the demographic shift of boomers moving towards retirement the population over age 65 will climb from just over 13 percent to nearly 22 percent. This shift will bring with it a progressive increase in attention to issues related to delaying retirement, retiree health availability and affordability, long term care, and how to produce income in retirement. There will be a tendency to rely upon generalizations or hypotheticals as opposed to detailed analysis. This report is intended to show the inherent risk in not undertaking detailed analysis for concerned parties.

Conclusion

133

X

References

135

136

Retirement Security in the United States

Alderson, Michael J., and Jack L. VanDerhei. “Disturbing the Balance in Corporate Pension Policy: The Case of Excess Asset Reversion Legislation.” Benefits Quarterly (Third Quarter, 1991): 38–51. Allen, Everett T., Joseph J. Melone, Jerry S. Rosenbloom, and Jack L. VanDerhei. Pension Planning: Pension, Profit-Sharing, and Other Deferred Compensation Plans. Eighth edition. Homewood, IL: Richard D. Irwin, Inc., 1997. Auer, Mary Jane. “A Mixed Bag for NationsBank.” Institutional Investor. Vol. 33, no. 1 (January 1999): 126. Bone, Christopher M. Statement before the ERISA Advisory Council Working Group on Hybrid Plans. U.S. Department of Labor, September 9, 1999. Campbell, Sharyn. “Hybrid Retirement Plans: The Retirement Income System Continues to Evolve.” EBRI Issue Brief no. 171 (Employee Benefit Research Institute, March 1996). Chambers, Robert G. Testimony on behalf of the Association of Private Pension and Welfare Plans before the ERISA Advisory Council Working Group on Hybrid Plans. Washington, DC: U.S. Department of Labor, September 9, 1999. Choi, James J., David Laibson, Brigitte C. Madrian, and Andrew Metrick. “Saving for Retirement on the Path of Least Resistance.” Originally prepared for Tax Policy and the Economy 2001. Updated draft: July 19, 2004. Chu, Kathy. “Rule would encourage automatic 401(k) enrollment.” http:// www.usatoday.com/money/perfi/retirement/2005-08-21-retirement-usat_ x.htm (last accessed, 4/17/2006). Copeland, Craig. “Employment-Based Retirement Plan Participation: Geographic Differences and Trends.” EBRI Issue Brief no. 274 (Employee Benefit Research Institute, October 2004). Copeland, Craig, Jack VanDerhei, and Dallas Salisbury. “Social Security Reform: Evaluating Current Proposals.” EBRI Issue Brief no. 210 (Employee Benefit Research Institute, June 1999). Demby, Elayne Robertson. “Cash Balance Converts.” Plan Sponsor (June 1999): 22–30. ________. “Cash Balance ‘Whipsaw.’” Plan Sponsor (May 1999): 74–76.

References

137

138

Retirement Security in the United States

Fidelity Investments. Building Futures, Volume V: How Workplace Savings Are Shaping the Future of Retirement. Boston, MA: Fidelity Investments, 2004. Gale, William G., J. Mark Iwry, and Peter R. Orszag. “The Automatic 401(k): A Simple Way to Strengthen Retirement Savings.” Retirement Security Project, March 2005. Gebhardtsbauer, Ron. “Testimony on Cash Balance Plans, Conversions and Disclosure.” For the U.S. Department of Labor ERISA Advisory Council, May 6 and June 9, 1999. Hewitt Associates. 2005 Trends and Experience in 401(k) Plans. Lincolnshire, IL: Hewitt Associates, LLC, 2005. Holden, Sarah, and Jack VanDerhei. “The Influence of Automatic Enrollment, Catch-Up, and IRA Contributions on 401(k) Accumulations at Retirement.” ICI Perspective, Vol. 11, no. 2, and EBRI Issue Brief no. 283 (Investment Company Institute and Employee Benefit Research Institute, July 2005). ________. “Can 401(k) Accumulations Generate Significant Income for Future Retirees?” ICI Perspective, Vol. 8, no. 3, and EBRI Issue Brief no. 251 (Investment Company Institute and Employee Benefit Research Institute, November 2002). ________. “Appendix: EBRI/ICI Accumulation Projection Model.” ICI Perspective. Vol. 8, no. 3A (Investment Company Institute, November 2002— Appendix). ________. “Contribution Behavior of 401(k) Plan Participants.” ICI Perspective. Vol. 7, no. 4, and EBRI Issue Brief no. 238 (Investment Company Institute and Employee Benefit Research Institute, October 2001). Ibbotson Associates. SBBI (Stocks, Bonds, Bills, and Inflation) 2002 Yearbook: Market Results for 1926–2001. Chicago, IL: Ibbotson Associates, Inc., 2002. ICMA-RC, “Automatic Enrollment.” Washington Perspective (Nov. 23, 2005), www.icmarc.org/xp/rc/plansponsor/regs/perspective/2005/11/ 200511automaticenrollment.html (last accessed: 2/23/06). Ippolito, Richard A. Pension Plans and Employee Performance: Evidence, Analysis and Policy. Chicago: University of Chicago Press, 1998. Madrian, Brigitte C., and Dennis F. Shea. “The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior.” NBER Working Paper. No. 7682. Cambridge, MA: National Bureau of Economic Research, May 2000. Mitchell, Olivia S. “Worker Knowledge of Pension Provisions.” Journal of Labor Economics. Vol. 6, no. 1 (January 1988): 21–40. Purcell, Patrick. “Automatic Enrollment in Section 401(k) Plans.” CRS Report for Congress (October 14, 2004).

Employee Benefit Research Institute

________. “Pension Issues: Cash-Balance Plans.” CRS Report for Congress (May 24, 1999). Quick, Carol. “An Overview of Cash Balance Plans.” EBRI Notes, no. 7 (Employee Benefit Research Institute, July 1999): 1–8. Rappaport, Anna M., Michael L. Young, Christopher A. Levell, and Brad A. Blalock. “Cash Balance Pension Plans.” In Michael Gordon, Olivia S. Mitchell, and Marc Twinney, eds., Positioning Pensions for the Twenty-First Century. Philadelphia: University of Pennsylvania Press, 1997. Sher, Larry. “A Workable Alternative to Defined Benefit Plans.” Contingencies (September/October 1999): 20–26. Thaler, Richard H., and Shlomo Benartzi. “Save More Tomorrow™: Using Behavioral Economics to Increase Employee Saving.” Journal of Political Economy. Vol. 112, no. 1, Pt. 2 (2004): S164–S187. Utkus, Stephen P. A Recent Successful Test of the SMarT Program. Valley Forge, PA: The Vanguard Center for Retirement Research and The Vanguard Group, November 2002. Toder, Eric et al. Modeling Income in the Near Term: Projections of Retirement Income Through 2020 for the 1931–1960 Birth Cohorts. Final Report SSA Contract No: 600-96-27332. Washington, DC: The Urban Institute, 1992. Towers Perrin. “The Controversy Over Cash Balance Plans.” Hot Topics (1999). VanDerhei, Jack. “Measuring Retirement Income Adequacy, Part One: Traditional Replacement Ratios and Results for Workers at Large Companies.” EBRI Notes, no. 9 (Employee Benefit Research Institute, September 2004): 1–12. ________. “The Recapture of Excess Pension Assets.” Benefits Quarterly. Vol. 1, no. 3 (Third Quarter 1985): 1–13. VanDerhei, Jack, and Craig Copeland. “Can America Afford Tomorrow’s Retirees: Results From the EBRI-ERF Retirement Security Projection Model.” EBRI Issue Brief no. 263 (Employee Benefit Research Institute, November 2003). ________. “The Changing Face of Private Retirement.” EBRI Issue Brief no. 232 (Employee Benefit Research Institute, April 2001). Vanguard Center for Retirement Research. Automatic Enrollment: Vanguard Client Experience. Valley Forge, PA: The Vanguard Group, July 2001. Warshawsky, Mark J. “Funding of Defined Benefit Pension Plans.” In Michael Gordon, Olivia S. Mitchell, and Marc Twinney, eds., Positioning Pensions for the Twenty-First Century. Philadelphia: University of Pennsylvania Press, 1997. Watson Wyatt. “Choosey Employees Choose Lump Sums!” Watson Wyatt Insider. Vol. 8, no. 4 (April 1998).

139

140

Retirement Security in the United States

Williamson, Christine. “Stock Allocation Hikes Predicted as Defined Benefit Plans Shift to Cash Balance.” Pensions & Investments (September 6, 1999): 1ff. Yakoboski, Paul J. “Male and Female Tenure Continues to Move in Opposite Directions.” EBRI Notes no. 2 (Employee Benefit Research Institute, February 1999): 1–4.

Appendix ESTIMATING CURRENT AND FUTURE ACCRUED BENEFITS AND ACCOUNT BALANCES FROM THE EBRI-ERF MODEL

141

T

he EBRI-ERF model is based on a six-year time series of administrative data from more than 10 million 401(k) participants and more than 30,000 plans, as well as a time series of several hundred plan descriptions used to provide a sample of the various defined benefit and defined contribution plan provisions applicable to plan participants. In addition, several public surveys based on participants’ self-reported answers (the Survey of Consumer Finances [SCF], the Current Population Survey [CPS], and the Survey of Income and Program Participation [SIPP]) were used to model participation, wages, and initial account balance information. This information is combined with U.S. Department of Labor Form 5500 data to model participation and initial account balance information for all defined contribution participants, as well as contribution behavior for non-401(k) defined contribution plans. Asset allocation information is based on previously published results of the EBRI/ICI ParticipantDirected Retirement Plan Data Collection Project and employee contribution behavior to 401(k) plans is provided by an expansion of a method based on both employee demographic information as well as plan matching provisions. A combination of Form 5500 data and self-reported results was also used to estimate defined benefit participation models; however, it appears information in the latter is rather unreliable with respect to estimating current and/or future accrued benefits. Therefore, a database of defined benefit plan provisions for salary-related plans was constructed to estimate benefit accruals. Combinations of self-reported results were used to initialize IRA accounts. Future IRA contributions were modeled from SIPP data, while future rollover activity was assumed to flow from future separation from employment in those cases in which the employee was participating in a defined contribution plan sponsored by the previous employer. Industry data are used to estimate the relative likelihood that the balances are rolled over to an IRA, left with the previous employer, transferred to a new employer, or used for other purposes. Appendix

143

144

Retirement Security in the United States

Defined Benefit Plans A stochastic job duration algorithm was estimated and applied to each individual in the EBRI-ERF model to predict the number of jobs held and age at each job change. Each time the individual starts a new job, the EBRI-ERF model simulates whether or not it will result in coverage in a defined benefit plan, a defined contribution plan, both, or neither. If coverage in a defined benefit plan is predicted, time series information from the Bureau of Labor Statistics (BLS) is used to predict what type of plan it will be.54 While the BLS information provides significant detail on the generosity parameters for defined benefit plans, preliminary analysis indicated that several of these provisions were likely to be highly correlated (especially for integrated plans). Therefore, a time series of several hundred defined benefit plans per year were coded to allow for assignment to the individuals in the EBRI-ERF model.55 Although the Tax Reform Act of 1986 at least partially modified the constraints on integrated pension plans by adding Sec. 401(l) to the Internal Revenue Code, it would appear that a significant percentage of defined benefit sponsors have retained primary insurance amount (PIA)offset plans. In order to estimate the offset provided under the plan formulae, the EBRI-ERF model computes the employee’s average indexed monthly earnings, primary insurance amount, and covered compensation values for the birth cohort.

Defined Contribution Plans Initial Account Balances Previous studies on the EBRI/ICI Participant-Directed Retirement Plan Data Collection Project have analyzed the average account balances for 401(k) participants by age and tenure.56 Recently published results show that the year-end 1998 average balance ranged from $4,479 for participants in their 20s with less than three years of tenure with their current employer to $198,595 for participants in their 60s who have been with the current employer for at least 30 years (thereby effectively eliminating any capability for IRA rollovers). Unfortunately, the EBRI/ICI database does not currently provide detailed information on other types of defined contribution plans nor does it allow analysis of defined contribution balances that may have been left with previous employers. The EBRI-ERF model uses self-reported

Employee Benefit Research Institute

responses for whether an individual has a defined contribution balance to estimate a participation model and the reported value is modeled as a function of age and tenure. Contribution Behavior Previous research on employee contribution behavior to 401(k) plans has often been limited by lack of adequate data. This is primarily due to the types of matching formulae utilized by sponsors. While these formulae are often complicated due to the desire of sponsors to provide sufficient incentives to non-highly compensated employees to contribute in order to comply with technical nondiscrimination testing, this complexity makes it virtually impossible to appropriately analyze the employee’s behavior if one is forced to observe either aggregate plan data or use information on the plan contribution formulae provided by the participant. With the exception of studies based on administrative data, employee contribution behavior is typically assumed to be a function of employee demographic data and perhaps an employee’s estimate of the employer matching rate or a proxy based on Form 5500 data. However, a significant percentage of the employee contribution behavior appears to be determined by plan-specific provisions. For example, the percentage of employees contributing up to the maximum amount of compensation matched, the 402(g) limit, or the plan maximum was studied by EBRI in 1996. It would appear that well over 50 percent of the employee contribution is explained by these “corner points” which would not be picked up in the data described above. Recently EBRI provided preliminary findings57 introducing new methodology to expand the usefulness of modeling these data, as well as a better understanding of contribution behavior by 401(k) plan participants. A sequential response regression model was used to allow for the differing incentives faced by the employees at various levels of contributions. Based on findings from 137 distinct matching formulae, EBRI estimated a behavioral model that is able to control for the tendency of employers to substitute between the amount they match per dollar of employee contribution and the maximum percentage of compensation they are willing to match. Employee contribution behavior was decomposed into a series of 1 percent of compensation intervals, therefore making it possible to model not only the marginal incentives to contribute at that interval but also the “option value” that making the contribution at that interval provides for the employee.

145

146

Retirement Security in the United States

Contribution behavior for defined contribution plans other than 401(k) plans is estimated from self-reported responses to public survey data. Investment Returns Although the EBRI-ERF model is designed to generate investment rates of return on a stochastic basis, this analysis presents the results obtained from running it in a deterministic mode. The same asset-specific rates of return were adopted that were used in the Social Security Administration’s Model of Income in the Near Term (MINT) model.58

Additional Assumptions for Retirement Income Projecting Social Security Benefits Three different Social Security scenarios are used to estimate the Social Security benefit levels of the total retirement income. Each of the scenarios under the intermediate assumptions within the 2000 Social Security Trustee’s Report meet the 75-year actuarial balance standard established by Congress. The estimates are generated from the SSASIM policy simulation model. This model is able to replicate actuarial balance and benefit estimates of the Social Security actuaries’ model. In addition, SSASIM allows for the changing of numerous policy parameters, including the normal retirement age, the PIA factors, and the Old Age and Survivors Insurance (OASI) and Disability Insurance (DI) tax rates. Given the projected financial condition of the Social Security program,59 we ran two reform scenarios designed to ensure 75-year solvency of the program. Under the first alternative, benefits were reduced.60 Under the second alternative, both the Social Security normal retirement age and the tax rates were increased.61 Housing Although it is not clear from the literature that retirees can be counted on to liquidate this source of wealth, we simulated whether households would be expected to have net housing equity at retirement, and if so, its expected value. Under the baseline scenario, it was assumed that retirees would not use their net housing equity to supplement their retirement income in any way (including housing equity loans). The second scenario assumed any net housing equity is annuitized at retirement. Given the stochastic nature of the analysis, a third scenario was modeled with the assumption that housing equity is not liquidated until the time it is

Employee Benefit Research Institute

first needed to mitigate an annual deficit. At that point it is assumed that any residual value is invested in the same manner as an individual account retirement plan.

147

Endnotes

149

1. The income sources are divided into Social Security (Old-Age, Survivors, and Disability Insurance (OASDI) payments), earnings, private pensions, public pensions, investment income, and other. Private pensions include survivor, disability, and retirement income pensions from corporate or union sponsors. In addition, regular payments from individual retirement accounts (IRAs), Keoghs, and 401(k)-type accounts are included in private pensions. Public pensions include payments from survivor, disability, and retirement income pensions from federal, U.S. military, and state or local sponsors. Investment income includes interest, dividend, and rent income. Other income includes unemployment compensation, workers’ compensation, supplemental security income, public assistance, veteran’s benefits, educational assistance, child support, alimony, financial assistance, other pensions, and other undefined income. Other pension income included in the other income category includes survivor payments from U.S. railroad retirement, workers’ compensation, Black Lung, regular payments from estates, trusts, annuities, or life insurance, and other survivor payments; disability payments from U.S. railroad retirement, accidental or disability insurance, Black Lung, workers’ compensation, state temporary sickness, and other disability payments; and retirement payments from U.S. railroad, regular payments from annuities or paid-up insurance policies, and other retirement payments. 2. See “Measuring Retirement Income Adequacy, Part One: Traditional Replacement Ratios and Results for Workers at Large Companies,” EBRI Notes, no. 9 (Employee Benefit Research Institute, September 2004): 2–12. 3. Recent work by Hewitt Associates has projected what employees working for large employers are likely to experience in terms of replacement ratios at retirement. Replacement ratios typically attempt to provide an indication of the percentage of income earned just prior to retirement that will be replaced in retirement. This typically involves a numerator that combines annuity payments from Social Security and defined benefit plans with an annuitized amount from defined contribution and IRAs. The denominator will be based on an average of final earnings just prior to retirement age. 4. Unlike many other models, the model used in this analysis does NOT merely assume that a retiree will survive to his or her average life expectancy. Unfortunately results generated under these assumptions would provide the amount necessary to pay for retirement expenditures only approximately 50

Endnotes

151

152

Retirement Security in the United States

percent of the time. Instead, this model considers the entire distribution of possible future lifetimes on a gender-specific basis and allows the concept of longevity risk to be explicitly modeled. This concept is explained in more detail in Chapter 4. 5. The Northeast region includes the states of Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, and Pennsylvania. The Midwest region includes the sates of Ohio, Indiana, Illinois, Michigan, Wisconsin, Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas. The South region includes the states of Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, Florida, Kentucky, Tennessee, Alabama, Mississippi, Arkansas, Louisiana, Oklahoma, and Texas; while the West region includes the states of Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, Nevada, Washington, Oregon, California, Alaska, and Hawaii. 6. The 2003 OASDI Trustees report subsequently reduced the assumed general inflation rate to 3.0 percent. The actuaries at the Center for Medicare & Medicaid Services developed a personal health care chain-type index that is a composite index of health care prices in the overall health care economy, which they predict will rise at a 3.5 percent level annually from 2004-2008 and 3.9 percent annually from 2009-2012. 7. While the medical consumer price index only accounts for the increases in prices of the health care services, it does not account for the changes in the number and/or intensity of services obtained. Thus, with increased longevity, the rate of health care expenditure growth will be significantly higher than the 4.0 percent medical inflation rate, as has been the case in recent years. 8. See the appendix for a description of how retirement income was estimated from Social Security and defined benefit plans and retirement wealth was estimated from defined contribution and cash balance plans as well as IRAs. Moreover, under some scenarios net housing equity was assumed to contribute to the retiree’s ability to pay for retirement expenditures. 9. For additional detail on how these findings differ by assumptions for the use of housing equity to pay retirement expenses as well as alternative approaches to Social Security reform, see EBRI Issue Brief no. 263, “Can America Afford Tomorrow’s Retirees: Results From the EBRI-ERF Retirement Security Projection Model” (Employee Benefit Research Institute, November 2003), available online at www.ebri.org). 10. See Holden and VanDerhei (September 2005) for more detail about these topics. All numbers from this section are from that study. 11. This chapter is partially based on Jack VanDerhei, “The Controversy of Traditional vs. Cash Balance Plans,” ACA Journal, Vol. 8, no. 4 (Fourth quarter 1999): 7–16.

Employee Benefit Research Institute

12. Although this chapter focuses exclusively on cash balance plans, hybrid arrangements that combine traditional defined benefit and defined contribution concepts include pension equity plans, age-weighted profit sharing plans, new comparability plans, floor-offset plans, new comparability profit-sharing plans and target plans (Campbell, 1996). 13. Jack L. VanDerhei, “Key Issues in the Cash Balance Debate,” Invited Testimony for the Senate Health, Education, Labor and Pensions Committee Hearing on Hybrid Pensions, September 1999. 14. A brief review of the Pension Protection Act of 2006 is provided at the end of the chapter. 15. All assumptions for this figure replicate those in Purcell (1999) with the exception of the benefit accrual rate which was decreased to 0.91 percent to allow for benefit equivalence of the two programs assuming 40 years of participation in the same program. The pay credits varied as follows: years 1–10: 4 percent, 11–20: 5.5 percent, 21–40: 7 percent. 16. For example, age-weighted pay credits under the cash balance plans and early retirement provisions under the final-average plan. 17. Note that they will not be exactly equal given that the pay credit differs from the assumed interest credited to the cash balance plan (5.6 percent). 18. In addition to these retirement plan-specific reasons, there may also be overall compensation or administrative concerns that are specifically addressed through a conversion. Two of the more common reasons include supporting a total compensation philosophy in the context of a new performance-based arrangement with employees, and providing a platform for merging disparate pension plans as a result of merger and acquisitions activity (Towers Perrin, 1999). 19. Using both administrative records and worker reports of pension provisions. 20. In the case of the job-changer, it is assumed that the full amount of any cash balance proceeds would be reinvested in a tax-deferred retirement savings account and earn an average annual rate of return of 8.65 percent, while the employee covered by a final-average plan would remain in a terminated vested status and not receive lump-sum distributions. 21. The calculation is obviously more complicated in an integrated plan. 22. For example, Eastman Kodak reportedly introduced a one-time match to its 401(k) plan to counterbalance losses from its conversion from a final average plan to a cash balance plan (Morrow, 1999).

153

154

Retirement Security in the United States

23. In addition to the potential cash flow problems arising from increased LSDs under cash balance plans, the liability durations of cash balance plans appear to be between seven to eight years as opposed to the 12- to 20-year durations typically calculated for traditional final average plans. Although the eventual impact (once the various transition provisions allow more of the liabilities to be generated via the new cash balance component) of the decreasing liability durations on the plan sponsor’s asset allocation is debatable (Williamson, 1999) it would appear that the expected rate of return on cash balance portfolios will remain significantly greater than the expected interest rate credited to the employees. 24. See Bone (1999) for a more complete description of the calculations required under FASB Statement No. 87. 25. Sher (1999, p. 22) reports that more than two-thirds of the plans included in the PricewaterhouseCoopers survey used an interest rate that was approximately equal to or less than 30-year Treasury bond rate at the time of the conversion. However some employers may desire to use a higher discount rate because the current 30-year Treasury bond rates are low relative to historical levels. The wear-away period actually experienced by a participant will be a function of the differential between the opening cash balance account and the present value of the accrued benefits under the previous defined benefit plan, as well as the future changes in discount rates. If the discount rate falls after the conversion, the present value of the previous benefits will increase, and the wear-away period experienced by the participant will increase (especially if the interest rate credited to the cash balance account is pegged to the 30-year Treasury bond rate). However, if the discount rate increases, the present value of the previous benefit will decrease, thereby reducing the wear-away period. 26. Dallas L. Salisbury, Statement before the Committee on Ways and Means Subcommittee on Oversight United States House of Representatives Hearing on Retirement Security and Defined Benefit Pension Plans, June 20, 2002. 27. A 1998 study by Watson Wyatt Worldwide found that 65 percent of defined benefit participants over age 60 elected a lump-sum distribution when it was available. The study is available on the Internet at www.watsonwyatt. com/us/pubs/insider/showarticle.asp?ArticleID=7249&Component=The+Insi der 28. Historically, some private employers had chosen to sponsor contributory defined benefit plans in which employees would need to contribute a specified amount (often a percentage of compensation) if they desired to participate in the plan. 29. These limits are in addition to the section 415(c) limits that apply generically to all defined contribution plans (the lesser of $44,000 or 100 percent of compensation in 2006).

Employee Benefit Research Institute

30. For employees age 50 or older, an opportunity to contribute an additional $5,000 per year on a before tax basis currently exists. However, this is one of the EGTRRA provisions that is scheduled to sunset early in the next decade unless Congress extends (perhaps permanently) this opportunity. 31. In general, employees earning more than $100,000 in 2006 would be considered to be highly compensated. This number is indexed for inflation in the future. 32. Technically, this may introduce further complexity; however, given that employee after-tax contributions (as well as employer matching contributions) are included in a similar test known as an ACP test. 33. This material first appeared in EBRI Issue Brief no. 273, “Health Savings Accounts and Other Account-Based Health Plans,” Employee Benefit Research Institute (September 2004). 34. The maximum annual contribution is actually the sum of the limits that are determined separately for each month. The monthly contribution limit is 1/12 of the lesser of the annual deductible or the maximum annual contribution. If an individual first becomes covered by a high-deductible health plan mid-year, the annual contribution limit is pro-rated, and the monthly contribution limit is based on the number of full months of eligibility. As an example, an individual who enrolled in a plan on July 1 with a $1,000 deductible would be eligible to contribute one-half (6/12) of the annual maximum contribution or $500 to the HSA. 35. Permitted insurance also includes worker’s compensation, tort liabilities, and liabilities related to ownership or the use of property (such as automobile insurance). 36. Only Medicare enrollees ages 65 and older are allowed to pay insurance premiums from an HSA. A Medicare enrollee under age 65 cannot use an HSA to pay insurance premiums. 37. The catch-up contribution is not indexed to inflation after 2009. 38. William G. Gale, “Tax Simplification: Issues and Options,” Based on testimony submitted to: Congress of the United States, House of Representatives, Committee on Ways and Means, Subcommittee on Oversight, Subcommittee on Select Revenue Estimates, July 17, 2001. 39. Growth and Investment Tax Plan would make Save at Work accounts “prepaid” or Roth-style. 40. Portions of this chapter previously draw heavily from material appearing in Sarah Holden and Jack VanDerhei, “The Influence of Automatic Enrollment, Catch-Up, and IRA Contributions on 401(k) Accumulations at Retirement,” EBRI Issue Brief and ICI Perspectives (July 2005).

155

156

Retirement Security in the United States

41. Vanguard (July 2001) also reports that most plan sponsors chose a default contribution rate of 3 percent or less (while about a quarter of plan sponsors selected a default contribution rate of 4 percent or higher). Hewitt (2005) reports that about a third of plan sponsors with automatic enrollment select a default contribution rate of 2 percent or less; about half choose 3 percent; and 17 percent select a default contribution rate of 4 percent or more. 42. Holden and VanDerhei (October 2001) find more than half of participants offered a match in 1999 were offered a match on up to at least 6 percent of salary or more. 43. Patrick Purcell, “Automatic Enrollment in Section 401(k) Plans,” CRS Report for Congress, October 14, 2004. 44. Chu (2006). 45. www.icmarc.org/xp/rc/plansponsor/regs/perspective/2005/11/ 200511automaticenrollment.html, last accessed: 4/17/2006. 46. Chu, op. cit. 47. Gale, Iwry, and Orszag (2005). 48. Choi et al. (2001) also study two other large companies’ 401(k) plans, in addition to the large health services company (that was initially analyzed in Madrian and Shea (May 2000)). Choi et al. (2004) consider 11 large companies with 401(k) plans implementing a variety of changes (e.g., automatic enrollment, eligibility rules, savings survey). 49. Note that these prices (which expresses the amount of lump sum that needs to be available at the time the immediate annuity is purchased) will be determined by, among other things, the life expectancies and the discount rates at the time of the future purchase. 50. See Holden and VanDerhei (September 2005). 51. This stylized example abstracts from any considerations that may be necessary to account for changes in tax rates over time and/or the possibility that the employee would make use of Roth 401(k) plan contributions. 52. This was the implicit assumption used in Mary Williams Walsh, “When Your Pension Is Frozen,” New York Times, Week in Review, January 22, 2006, Section 4, p. 3. 53. Two other plans that were not modeled in this study are likely to have an indemnification contribution rate between final-/career-average plans and cash balance plans. Flat benefit plans provide an annual income based exclusively on the number of years of participation (compensation is ignored) and pension equity plans are a type of hybrid plan that provides benefits that can be more valuable to older and/or longer tenured employees than cash balance plans in general.

Employee Benefit Research Institute

54. The model is currently programmed to allow the employee to participate in a nonintegrated career average plan; an integrated career average plan; a five-year final-average plan without integration; a three-year final average plan without integration; a five-year final average plan with covered compensation as the integration level; a three-year final-average plan with covered compensation as the integration level; a five-year final-average plan with a PIA offset; a three-year final-average plan with a PIA offset; a cash balance plan, or a flat benefit plan 55. BLS information was utilized to code the distribution of generosity parameters for flat benefit plans. 56. Holden and VanDerhei (2001). 57.VanDerhei and Copeland (2001). 58. MINT assumes a CPI growth rate of 3.50 percent, a real rate of return for stocks of 6.98 percent, and a real rate of return for bonds of 3.00 percent. It subtracts 1 percent from each of the stock and bond real rates of return to reflect administrative cost (Toder et al., 1999). 59. The present Social Security program has been shown to be financially unsustainable in the future without modification to the current program. For additional detail, see Copeland, VanDerhei, and Salisbury (1999). 60. This scenario involves gradually reducing the benefits of those starting to receive retirement and survivor’s benefits. The reduction starts immediately and reaches 10 percent of present law benefits in 2010, 15 percent in 2016, and 22 percent in 2022. 61. Under this reform alternative, the normal retirement age continues its increase from 65 to 67 but at a faster pace than under current law. Thereafter, the normal retirement age is indexed to longevity (currently assumed to be one month every two years). An increase from 10.6 percent to 12.35 percent in 2030 and to 13.50 percent in 2050 of the OASI tax rate completes the proposal.

157

2121 K Street, NW Suite 600 Washington, DC 20037-1896 202.659.0670

ISBN 0-86643-100-4 5 1 9 9 9>

9 780866 431002