ASSESSING SOCIAL SECURITY REFORM ALTERNATIVES · i EBRI EMPLOYEE BENEFIT RESEARCH INSTITUTE®...

i

EBRIEMPLOYEE

BENEFIT

RESEARCH

INSTITUTE ®

ASSESSING SOCIAL SECURITY REFORM ALTERNATIVES

An EBRI ERF Policy Forum

Edited by Dallas L. Salisbury

Assessing Social Security Reform Alternatives

ii

© 1997 Employee Benefit Research InstituteEducation and Research Fund2121 K Street, NW, Suite 600Washington, DC 20037-1896(202) 657-0670web site www.ebri.org

All rights reserved. No part of this publication may be used or reproduced in any manner whatsoeverwithout permission in writing from the Employee Benefit Research Institute Education and ResearchFund except in the case of brief quotations embodied in news articles, critical articles, or reviews. Theviews expressed in this publication should not be ascribed to the officers, trustees, members, or othersponsors of the Employee Benefit Research Institute, the EBRI Education and Research Fund, or theirstaffs. Nothing herein is to be construed as an attempt to aid or hinder the adoption of any pendinglegislation, regulation, or interpretative rule, or as legal, accounting, actuarial, or other such profes-sional advice.

Library of Congress Cataloging-in-Publication Data

Assessing social security reform alternatives / edited by Dallas L. Salisburyp. cm.

Includes bibliographical references (p. )ISBN 0-86643-087-31. Social security—United States. I. Salisbury, Dallas L.

HD7125.A783 1997368.4'3'00973—dc21 97-13587

CIP

iii

Th

e p

rem

ier

orga

niz

atio

n

EBRIEMPLOYEE

BENEFIT

RESEARCH

INSTITUTE

Established in 1978, the Employee BenefitResearch Institute (EBRI) is the only nonprofit,nonpartisan organization committed to originalpublic policy research and education on economicsecurity and employee benefits.

EBRI does not lobby or endorse specific ap-proaches. Rather, it provides balanced analysis ofalternatives based on the facts. Through its activi-ties, EBRI is able to fulfill its mission.

Since its inception, EBRI’s membership hasgrown to represent a cross section of pensionfunds; businesses; trade associations; labor unions;health care providers and insurers; governmentorganizations; and service firms, including actu-arial firms, employee benefit consulting firms, lawfirms, accounting firms, and investment manage-ment firms.

Today, EBRI is recognized as one of the mostauthoritative and objective resources in the nationon employee benefit issues—health care, pensions,and economic security.

for employee benefits research


iv

Table of Contents

List of Tables and Charts .................................... vii

Preface ................................................................. viii

About the Authors ................................................ ix

Executive Summary: Assessing Social SecurityReform AlternativesChristopher R. Conte

Introduction ..................................................... 1The Context: Pitfalls and Possibilities ........... 1The Mortality Debate ...................................... 2Predicting Rates of Return .............................. 3Macroeconomic Effects .................................... 4The Perception Gap ......................................... 5Reform Alternatives......................................... 6What’s Next? .................................................... 8

1. A Framework for Analyzing and ComparingSocial Security PoliciesKelly Olsen, Jack VanDerhei, andDallas L. Salisbury

OASDI’s Finance Issues .................................. 9Policy Evaluation ............................................ 11Underlying Benefits........................................ 11Program Parameters ..................................... 12

Basic Plan Design...................................... 12Program Finance ........................................ 13

Comparing Outcomes .................................... 15Adequacy ..................................................... 16Equity .......................................................... 17Program-Specific Monetary Cost

Considerations ........................................ 18Extra-Programmatic Economic

Considerations ........................................ 19Effects on the Rest of the U.S. Retirement

System ..................................................... 23Governmental Effects ................................. 24Administrative Considerations.................. 25Political Considerations ............................. 25

Social Considerations ................................. 27Protection Against Uncertainties .............. 27Weighing Considerations to Determine

the Best Social Security Policy ............... 28Conclusion ...................................................... 28References ...................................................... 28

2. Economic Security: An Overview of SocialSecurityDavid V. Bryce and Robert B. Friedland

Introduction ................................................... 29Enacting the Social Security Act of 1935 ...... 29

The Impact of the Great Depression ......... 31The Influence of the Townsend

Movement ................................................ 32The Approach of the Committee on

Economic Security ...................................... 32The Social Security Act .............................. 33

The Debates Over Contributory Taxes ......... 34Contributory Taxes, the Initial CES

Report, and Tax Rate Adjustments ........ 34Opposition to Contributory Financing

and the Trust Fund ................................. 35The Reserve Fund and the 1939

Amendments ............................................... 37Retirement Age and Related Issues .............. 38The Maturing of Social Security ................... 39Bibliography ................................................... 41

3. EBRI Social Security Reform Analysis ProjectProgress Report: Phases 1 and 2Martin R. Holmer

Preface ............................................................ 43Understanding via Quantitative Analysis.... 43Rationale for a New Kind of Model ............... 43

EBRI-SSASIM2 Model Overview .............. 44Examining the Nature of the Problem ......... 45

Future Cost of Current-Law BenefitPolicy ....................................................... 45

Demographic Roots of the Cost Problem ... 46

v

Cost Uncertainty Using Trustees’ Views .. 46Cost Uncertainty Using Census’

Mortality View......................................... 49Size and Timing of the Cost Problem ........ 51

Examining Alternative Solutions.................. 52Current Benefits and Increased

Taxes (CBIT) ........................................... 53Reduced Benefits and Current

Taxes (RBCT) .......................................... 53DC Accounts and Reduced DB Benefits

(IARB) ...................................................... 55

4. Comments on the EBRI-SSASIM2 Modelas of January 1997Steve Goss

Introduction ................................................... 57Examining the Numbers ............................... 57Mortality Improvement and Fertility

Rates ........................................................... 58Effects of Simplification ................................ 59Developing Policy Alternatives ..................... 59Macroeconomic Implications ......................... 59Policy Performance Measures ....................... 59Conclusion ...................................................... 60

5. Social Security Reform: Beyond the ModelsMarilyn Moon

Introduction ................................................... 61Assumptions and Projections ........................ 61Issues to Consider .......................................... 62Conclusion ...................................................... 62

6. The Importance of Systematic AnalysisWilliam Beeman

Introduction ................................................... 63Examining the Baseline Projections ............. 63Handling Risk and Uncertainty .................... 63Conclusions .................................................... 64

7. Can America Afford to Grow Old?Gary Burtless

Introduction ................................................... 65Proposed Reforms .......................................... 66The Model ....................................................... 66Obtaining Symmetry ..................................... 66The Questions for Generation X ................... 66Cost-of-Living Calculations ........................... 67

8. A Comment on Mortality Assumptions,Low Wealth, and the Need for More SavingJames P. Smith

Introduction ................................................... 69Mortality Assumptions .................................. 69Mortality and Socio-Economic Status........... 70Achieving Fiscal Balance ............................... 70Conclusion ...................................................... 70

9. What Can We Learn from ModelingSocial Security?Gene Steuerle

Introduction ................................................... 73Allocating Resources ...................................... 73Increasing Savings ......................................... 74The Work Force .............................................. 74Retirement Age .............................................. 74The Impact of Government ........................... 75Integrating Private Plans in the Models ...... 75The Question of Risks .................................... 75Conclusion ...................................................... 76

10. Issues Concerning the ModelRobert Shapiro

Introduction ................................................... 77The Economy .................................................. 77Trust Fund Assets .......................................... 77Modeling Future Performance ...................... 78Conclusion ...................................................... 78

11. Social Security Reform: Gaps in PerceptionSusan Dentzer

Introduction ................................................... 79Social Security Participants as Investors ..... 79Annuities Markets ......................................... 80Raising Taxes ................................................. 80The Retirement Age ....................................... 81Medicare and Social Security ........................ 81

12. Investment Implications of Social SecurityReform: More Questions than AnswersGirard Miller

Introduction ................................................... 83Workers as Investors ..................................... 83Costs ............................................................... 84Administration ............................................... 85Regulation ...................................................... 85Macroeconomic Considerations ..................... 86Conclusion ...................................................... 86


vi

17. A Retirement System for the Future: PersonalSecurity AccountsAnn L. Combs

Introduction ................................................. 107The Personal Savings Account Plan ........... 107The Flat Benefit ........................................... 108Transition Costs ........................................... 109Risk ............................................................... 109The Government as Investor ........................110Conclusion .....................................................110

18. The Case for Incremental ChangeStanford G. Ross

Introduction .................................................. 111A Case for the Voluntary Approach ............. 111Conclusion .....................................................112

AppendicesA. Keeping Track of Social Security Reform

Proposals: A SummaryKelly Olsen

Introduction ..................................................113Reform Proposals ..........................................113

B. Daring to Touch the Third RailPaul Yakoboski

Introduction ................................................. 123A General Lack of Confidence ..................... 123The Public Views Reform ............................ 123

Benefit Cuts and Tax Increases ............... 123Trust Fund Equity Investment................ 125Individual Accounts .................................. 125

Implications.................................................. 125

C. Public Attitudes on Social Security ReformPamela Ostuw

Introduction ................................................. 127The Current System .................................... 127Reform Proposals ......................................... 127Changes in the Level of Benefits

Received .................................................... 128Expected Returns......................................... 128

D. Additional Information

Social Security Reform Bibliography .......... 131Internet Sites with Information on

Social Security Reform ............................. 135

E. Policy Forum Attendees List ....................... 137

13. Social Security: Myths and RealitiesRobert J. Myers

Introduction ................................................... 87Myths about the System................................ 87Modeling ......................................................... 89Privatization .................................................. 89Maintaining the System ................................ 89Conclusion ...................................................... 90

14. The Maintain Benefits Plan for Social SecurityRobert M. Ball

Introduction ................................................... 91The Problems with Individual Accounts ....... 91Administrative Problems .............................. 92Bringing the Current System into Balance .. 93Investing Reserves ......................................... 93Conclusion ...................................................... 95

15. The Case for the Individual Accounts OptionLawrence H. Thompson.

Introduction ................................................... 97Different Approaches ..................................... 97Providing for Individual Accounts ................ 98Risks Associated with Retirement

Planning .................................................. 98Cohort Life Expectancy .............................. 98Future Economic Growth and

Investment Returns ................................ 98Relative Earnings Level ............................. 98Individual Life Expectancy ........................ 99Inflation ...................................................... 99

How Risks Are Distributed in DifferentRetirement Income Vehicles ...................... 99

Preserving a Mixed System ........................... 99Administrative Efficiency .............................. 99Increased Personal Savings ........................ 100Preserving Medicare Financing .................. 101Conclusion .................................................... 101

16. Comments on the National Thrift PlanNeil Howe

Introduction ................................................. 103The National Thrift Plan ............................. 103Questions about Modeling ........................... 104Conclusion .................................................... 105

vii

List of Tables and Charts

TablesTable 1.1 Reasons for Changing Trust Fund

Balances, 1983–1996 ..................................... 10Table 3.1 Summary Estimates of Cost Problem

for Current-Law Benefit Policy underTrustees’ Report Intermediate-CostAssumptions ................................................... 46

Table 3.2 Summary Estimates of Cost Problemsfor Current-Law Benefit Policy AssumingShorter Life Expectancies ............................. 47

Table 3.3 Range of Cost Estimates forCurrent-Law Benefit Policy, Using ThreeSets of Trustees’ Report Assumptions .......... 47

Table 3.4 Alternative Estimates of PopulationAges 85 and Older in 2050 ............................ 49

Table 3.5 Effects of Generic Reforms on OASIProgram and on Economy ............................. 54

Table 3.6 Effects of Generic Reform on 1976Birth Cohort ................................................... 55

Table 14.1 Reducing Social Security’sLong-Term Deficit .......................................... 93

Appendix A, Table 1 A Point-by-PointComparison of Social Security ReformPlans ..............................................................115

ChartsChart 3.1 Cost Problem for Current-Law

Benefit Policy under Trustees’ ReportIntermediate-Cost Assumptions ................... 45

Chart 3.2 Historical and Projected LifeExpectancy at Birth under Trustees’Report Intermediate Cost Assumptions ....... 46

Chart 3.3 Long-Run Average and 2070 CostRate Distribution under StochasticAssumptions Consistent with Trustees’Report ............................................................. 48

Chart 3.4 Long-Run Average and 2070Actuarial Deficit Distribution underStochastic Assumptions Consistent withTrustees’ Report ............................................. 48

Chart 3.5 Distribution of Life Expectancy atBirth in 2070 under Stochastic AssumptionsConsistent with Trustees’ Report andApproximating Census Bureau’s .................. 50

Chart 3.6 Long-Run Average and 2070 CostRate Distribution under StochasticAssumptions Consistent with Trustees’Report Except for Use of Mortality DeclineAssumptions Approximating CensusBureau’s .......................................................... 50

Chart 3.7 Long-Run Average and 2070Actuarial Deficit Distribution underStochastic Assumptions Consistent withTrustees’ Report Except for Use of MortalityDecline Assumptions Approximating CensusBureau’s .......................................................... 51

Chart 3.8 Maximum-Fund-Ratio YearDistribution and Fund-Insolvency YearDistribution under Stochastic AssumptionsConsistent with Trustees’ Report Except forUse of Mortality Decline AssumptionsApproximating Census Bureau’s .................. 52

Chart 12.1 Asset Mix, 1990–1996 ........................ 84Chart 12.2 Participants’ Contribution,

1990–1996 ...................................................... 85Chart 14.1 Money’s Worth Ratios—Total

Composite Worker Comparison ofPL PAYGO, MB PAYGO, and MB withInvestment in Equities .................................. 94

Appendix B, Chart 1 Retirees’ AttitudesRegarding Social Security Reform/Workers’Attitudes Regarding Social SecurityReform .......................................................... 124

Appendix C, Chart 1 Percentage of RespondentsWho Agreed or Strongly Agreed,March 1995 ................................................... 128

Appendix C, Chart 2 Payroll Tax IncreasePreferences, March 1995 ............................. 129


viii

For decades, most retirees have viewed SocialSecurity as a vital program. This is largely due tothe fact that individuals have not saved enoughthemselves, or planned for retirement. The 1996Retirement Confidence Survey found that less thanone-third of retirees had done income or expenseplanning prior to actual retirement. Most did noteven know what their Social Security benefit wouldbe.

The papers in this volume provide a picture ofthe vital nature of the Social Security program. Atthe same time, they depict the long-term financingissues faced by the program due to changingdemographics. The papers highlight the high levelof public support the Social Security systemreceives, and current public concerns about theprogram’s future.

The papers set forth a framework for: analysisof reform options; details of work done by theEmployee Benefit Research Institute (EBRI) tofacilitate analysis and commentaries on its work;details on the various reform proposals setting thestage for current public policy debate; and assess-ments of the implications for the future.

This policy forum was the 40th in a seriesbegun in 1979. Our mission: to contribute to, toencourage, and to enhance the development ofsound employee benefit programs and sound publicpolicy through objective research and education.The EBRI Social Security Reform Project is amulti-year effort to ensure the availability ofobjective research tools as the nation grapples withthe future of this vital “employee benefit” program.

I wish to thank the contributors and thelarger group of forum participants for making the

session stimulating and insightful. I also wish tothank EBRI Members for making the forumpossible, and those who have made special grantsfor the Social Security Project: American Compen-sation Association, Amoco Foundation, AT&T,AT&T/Actuarial Sciences Associates, Inc., BarclaysGlobal Investors, CIGNA Corporation, CitibankN.A., Goldman Sachs & Co., Hewitt, ICMA Retire-ment Corporation, Investment Company Institute,Milliman & Robertson, Inc., Mobil Oil Corporation,National Taxpayers Union Foundation, PacificaTelesis Group, Principal Financial Group, SBCCommunications, Inc., State Street Bank and TrustCo., TIAA-CREF, United Food & CommercialWorkers Union, William H. Donner Foundation,and Zurich Kemper Investments, Inc.

This publication would not have been possiblewithout the work of Bill Pierron and Pam Ostuw,who administered the forum; the noble efforts ofDeborah Holmes, Lynn Miller, Maureen Richmond,and Cindy O’Connor in copy editing, proofing, andpublishing the volume; and all the members of theEBRI team for providing many forms of support. Inthe end, however, I take sole responsibility for anyerrors or omissions.

The views expressed in this book are solelythose of the authors and participants. They shouldnot be attributed to officers, trustees, EBRI Mem-bers, its staff, or its Education and Research Fund.In publishing this book, EBRI-ERF is making noeffort to influence any specific legislation. Com-ments on the contents and suggestions for workthat we might undertake in the future are encour-aged.

Dallas L. SalisburyEBRI President

May 1997

Preface

ix

About the Authors

Robert M. BallRobert M. BallRobert M. BallRobert M. BallRobert M. Ball is a consultant on Social Security andhealth and welfare policy to many organizationsand elected officials. From April 1962 to March1973, Ball was U.S. Commissioner of Social Secu-rity, serving under Presidents Kennedy, Johnson,and Nixon and was a member of the NationalCommittee on Social Security Reform during 1982and 1983. After some 30 years of service at theSocial Security Administration, Ball served as asenior scholar at the Institute of Medicine for sevenyears, and later as a visiting scholar at the Centerfor the Study of Social Policy. He is the author ofmany articles, books, and reports, includingPensions in the United States, Social Security:Today and Tomorrow, and Because We’re All in ThisTogether. He is a member of the current AdvisoryCouncil on Social Security and has served as bothmember and staff on other Social Security AdvisoryCouncils, including those in 1948 and 1991. Ballreceived his M.A. from Wesleyan University.

William BeemanWilliam BeemanWilliam BeemanWilliam BeemanWilliam Beeman is vice president and director ofeconomic studies for the Committee for EconomicDevelopment (CED), a nonprofit research andeducational organization of business and academicleaders devoted to the study of public policy prob-lems. Since joining the CED in 1987, he hasprepared numerous CED policy statements dealingwith national economic policies, including the 1995statement on pension policy entitled Who Will Payfor Your Retirement? The Looming Crisis. Cur-rently, he is project director for an ongoing CEDstudy on reforming the Social Security retirementsystem. Before joining the CED, Beeman wasassistant director of the Congressional BudgetOffice, where he headed CBO’s Fiscal AnalysisDivision. Prior to joining the CBO, Beeman waswith the Federal Reserve Board.

David BryceDavid BryceDavid BryceDavid BryceDavid Bryce is a research assistant at the NationalAcademy on Aging. He received his A.B. from theUniversity of Michigan and his M.S. from theUniversity of Texas.

Gary BurtlessGary BurtlessGary BurtlessGary BurtlessGary Burtless is a senior fellow in the EconomicsStudies Program at the Brookings Institution,where he does research on labor markets, incomeredistribution, and the economic effects of taxes.Before coming to Brookings in 1981, Burtlessserved as an economist in the Policy and Evalua-tion offices of the Secretaries of Labor and Health,Education and Welfare. He has also served asvisiting professor of public affairs at the Universityof Maryland. He has written and published numer-ous articles on applied econometrics andmicroeconomics. With Barry Bosworth and HenryAaron, he is the author of Can America Afford toGrow Old? Paying for Social Security; with MartinNeil Baily and Robert E. Litan he is the author ofGrowth With Equity: Economic Policymaking forthe Next Century. With Daniel Friedlander, hewrote Five Years After: The Long-Term Effects ofWelfare-to-Work Programs. He received his Ph.D. ineconomics from the Massachusetts Institute ofTechnology.

Ann L. CombsAnn L. CombsAnn L. CombsAnn L. CombsAnn L. Combs is a principal with William M.Mercer, Inc.’s Washington Resource Group. For-merly, she was the deputy assistant secretary forpolicy for the Pension and Welfare Benefits Admin-istration of the U.S. Department of Labor. Shepreviously served as senior legislative officer in theLabor Department’s Office of Congressional Affairs.Prior to joining DoL, she was a tax consultant atPrice Waterhouse and a senior analyst in pensionand labor policy at the Government ResearchCorporation. She worked for the National Associa-tion of Manufacturers, where she served asAlexander Trowbridge’s assistant on the NationalCommission on Social Security Reform, as associatedirector of labor relations, and as a programanalyst in employee benefits. She is a member ofthe current Advisory Council on Social Security.Combs received her J.D. from the George Washing-ton University.


x

Christopher R. ConteChristopher R. ConteChristopher R. ConteChristopher R. ConteChristopher R. Conte is a freelance writer based inthe Washington, DC, area. An EBRI fellow since1995, he has written on a wide range of socialpolicy issues, including Social Security, Medicare,and welfare reform, for such publications asGoverning Magazine, Congressional Quarterly, andthe AARP Bulletin. He also has reported exten-sively on telecommunications policy and the socialimplications of new information technologies for theBenton Foundation. In addition, he has writtenmajor articles about journalism. From 1979 to1995, Mr. Conte worked in the Washington bureauof the Wall Street Journal, where he coveredeconomic policy, banking, and transportation, andedited both foreign and domestic policy coverage.He also wrote the Washington Wire, a weeklycolumn on government and politics, and the LaborLetter, a column on labor and work place issues.Earlier, he worked as a reporter for the Congres-sional Quarterly and the Rutland Herald in Ver-mont. He holds a B.A. from Harvard.

Susan DentzerSusan DentzerSusan DentzerSusan DentzerSusan Dentzer is chief economics correspondent andthe economics columnist for U.S. News and WorldReport. Based in Washington, DC, where themagazine is headquartered, she writes about awide array of economic issues ranging from globalcompetition to health care reform. She specializesin writing about political-economic concerns, suchas U.S. fiscal and monetary policy; internationaleconomic issues, including U.S.-Japan relations;and domestic policy concerns such as the growth ofMedicare and other entitlement programs. In hercolumn, “On the Economy,” Dentzer has exploredsuch topics as the Social Security trust funds andthe outcome of recent trade disputes with Japan.She is a 1977 graduate of Dartmouth College,magna cum laude. Dentzer is also a member of theboard of Dartmouth, holds and honorary master ofarts degree from the institution, and is a recipientof the college’s Presidential Medal for Achievement.

Robert FriedlandRobert FriedlandRobert FriedlandRobert FriedlandRobert Friedland is the director of the NationalAcademy on Aging. Previously, he was researchdirector for the National Academy of Social Insur-ance. His public policy experience includes chiefeconomist for Maryland’s Medicaid program; seniorresearch associate at EBRI; director of the PublicPolicy Institute of the American Association ofRetired Persons; senior policy analyst with Project

HOPE’s Center for Health Affairs; and economiston the staff of the U.S. Bipartisan Commission onComprehensive Health Care, popularly known asthe Pepper Commission. He has written on issuespertaining to the financing and delivery of healthcare, long-term care, and retirement incomesecurity. His book, Facing the Costs of Long-TermCare, received the 1992 Elizur Write Award fromthe American Risk and Insurance Association.Friedland received his Ph.D. in economics from theGeorge Washington University.

Stephen GossStephen GossStephen GossStephen GossStephen Goss is deputy chief actuary in the Office ofthe Actuary at the Social Security Administration.He has been with the SSA since graduating fromcollege in 1973. Goss has been a staff participantrepresenting the Office of the Actuary at thePresident’s Commission on Pension Policy, the 1979and 1991 Advisory Councils, the National Commis-sion on Social Security, and the National Commis-sion on Social Security Reform. He has presentedpapers at the Society of Actuaries, the NationalAssociation of Insurance Commissioners, NationalConference on Private Long-Term Care Insurance,the American Academy of Actuaries, and theAtlantic Economic Society. He is an associate in theSociety of Actuaries. Goss received his M.S. inmathematics from the University of Virginia.

Martin R. HolmerMartin R. HolmerMartin R. HolmerMartin R. HolmerMartin R. Holmer, president of the Policy SimulationGroup, specializes in the design, implementation,and use of micro-sample and cell-based policysimulation models for risk assessment and riskmanagement in the pension, health insurance, andfinancial portfolio areas for both private-sector andpublic-sector clients. He has been director of incomesecurity policy research at the U.S. Department ofHealth and Human Services and vice president forasset/liability strategy at Fannie Mae, among otherpast positions. He has a Ph.D. in economics fromthe Massachusetts Institute of Technology.

Neil HoweNeil HoweNeil HoweNeil HoweNeil Howe is an author, historian, economist, anddemographer. He has served as a magazine editor,a foundation program officer, and a researchdirector; he is now a marketing, personnel, andgovernment affairs consultant to a variety ofcorporate and nonprofit clients. His current titlesinclude: Senior Advisor on Public Policy to theBlackstone Group, Chief Economist for the Na-

xi

tional Taxpayers Union Foundation, and SeniorAnalyst for the Concord Coalition. Howe haswritten extensively on budget policy and aging, onattitudes toward economic growth and socialprogress, on the collective personalities of Americangenerations, and on how generations succeed or failin creating endowments for their heirs. He coau-thored On Borrowed Time (with Peter G. Peterson),Generations (with William Strauss) and 13th GEN(also with William Strauss). He took his B.A. at theUniversity of California at Berkeley, studied abroadin France and Germany, and later received gradu-ate degrees in economics (M.A.) and history(M.Phil. with honors) from Yale University.

Girard MillerGirard MillerGirard MillerGirard MillerGirard Miller is the president and chief executiveofficer of the ICMA Retirement Corporation (RC), anonprofit corporation based in Washington, DC,serving 4,500 employers and 250,000 participants.ICMA Retirement Corporation provides investmentservices for $6 billion of retirement plan funds andplan administration for state and local govern-ments nationwide. Previously, Miller was seniorvice president and the head of the governmentalservices division at Fidelity Investments Institu-tional Services Company in Boston, MA. From 1981to 1987, Miller was director of the TechnicalServices Center of the Government Finance Offic-ers’ Association (GFOA). He founded the GFOA’sinvestment newsletter, Public Investor, and is theauthor of numerous GFOA publications, includingInvesting Public Funds and Pension Fund Invest-ing. Miller holds a master’s degree in publicadministration from the Maxwell School of PublicAffairs of Syracuse University and a master’sdegree in economics from Wayne State Universityin Detroit. He is a chartered financial analyst(CFA).

Marilyn MoonMarilyn MoonMarilyn MoonMarilyn MoonMarilyn Moon is a senior fellow in the Health PolicyCenter of the Urban Institute and public trustee ofthe Social Security and Medicare trust funds. Inthe fall of 1989, she served as a consultant to thePepper Commission. From 1986 to 1989, Moon wasdirector of the Public Policy Institute at the Ameri-can Association of Retired Persons. Prior to that,she worked as a senior research associate at theUrban Institute, a senior research analyst at theCongressional Budget Office, and as an associateprofessor of economics at the University of Wiscon-

sin-Milwaukee. She has written extensively onMedicare, poverty, health, income distribution, andlong-term care issues. Her recent publicationsinclude Medicare Now and in the Future, CanStates Take the Lead in Health Care Reform? andWomen and Long-Term Care. She also writes acolumn for the Washington Post, answering ques-tions on health care coverage. Moon received herPh.D. in economics from the University ofWisconsin.

Robert J. MyersRobert J. MyersRobert J. MyersRobert J. MyersRobert J. Myers is a member of the ProspectivePayment Assessment Commission and is anactuarial consultant to several foreign countries.He was chief actuary for the Social SecurityAdministration from 1947 to 1970, deputy commis-sioner of Social Security from 1981 to 1982, execu-tive director of the National Commission on SocialSecurity Reform from 1982 to 1983, chairman ofthe Commission on Railroad Retirement Reformfrom 1988 to 1990, and a member of the Commis-sion on the Social Security “Notch” Issue from 1993to 1994. He is the president of the InternationalFisheries Commission Pension Society and aconsultant on Social Security to William M. Mercer,Inc., the National Association of Life Underwritersand the Seniors’ Coalition. Myers has written fivebooks and more than 900 articles. Myers receivedhis LL.D. from Muhlenberg College and LehighUniversity and his M.S. from the University ofIowa.

Kelly OlsenKelly OlsenKelly OlsenKelly OlsenKelly Olsen is a research analyst with the EmployeeBenefit Research Institute, where her work focuseslargely on Social Security, employment-basedpensions, and income of the elderly. She hasworked as a Herman (Red) Somers intern at theNational Academy of Social Insurance and as aresearch assistant for computer and researchcourses at the Boston College Graduate School ofSocial Work (BCGSSW). She has worked directlywith the elderly in a variety of clinical and volun-teer settings and has had policy exposure throughlegislative internships, as an undergraduatepolitical science major, and as a student of socialwelfare policy at BCGSSW. She received a B.A. inMay 1993 from the University of Rochester with adouble major in political science and philosophy,and an M.S.W. in clinical gerontology from BostonCollege Graduate School of Social Work in May1996.


xii

Pamela OstuwPamela OstuwPamela OstuwPamela OstuwPamela Ostuw joined EBRI as a research/memberservices assistant in May 1996. At EBRI, sheresearches and writes on health care and pensiontopics, in addition to her involvement with theEBRI Social Security Reform Evaluation Projectand the EBRI/ASEC/Greenwald RetirementConfidence Survey. She is also the coordinator ofEBRI-ERF policy forums and the EBRI LillywhiteAward, as well as the EBRI Fellows ProgramManager. Previously, Pam was employed atGallaudet Research Institute in the Center forAssessment and Demographic Studies. She re-ceived her B.A. from Kenyon College, with honorsin psychology.

Stanford G. RossStanford G. RossStanford G. RossStanford G. RossStanford G. Ross is a senior partner in the law firmof Arnold and Porter in Washington, DC. He hasdealt extensively with public policy issues whileserving in the U.S. Treasury Department, on theWhite House domestic policy staff, as Commis-sioner of Social Security, and as a public trustee ofthe Social Security and Medicare trust funds. Hehas taught law at New York University, Harvard,Virginia and Georgetown Law Schools, and hasbeen a visiting fellow at the Hoover Institution,Stanford University. Ross has served as chairmanof the American Bar Association Tax SectionCommittee on Social Security and Payroll TaxProblems. Under the auspices of the IMF, WorldBank, OECD, and U.S. Treasury Department, hehas provided technical assistance on Social Securityand tax issues to various foreign countries. Hereceived his J.D. from Harvard Law School.

Dallas L. SalisburyDallas L. SalisburyDallas L. SalisburyDallas L. SalisburyDallas L. Salisbury is president and CEO of theEmployee Benefit Research Institute (EBRI),Washington, DC. He is also chairman and CEO ofthe American Savings Education Council (ASEC), apartnership of public- and private-sector institu-tions that undertakes initiatives to raise publicawareness about what is needed to ensure long-term personal financial independence. He iscurrently a member of the Board of Directors of theNational Academy of Social Insurance, the Board ofDirectors of The Health Project, and the AdvisoryBoard of the National Academy on Aging. He serveson many editorial advisory boards, including thoseof Employee Benefit News, Benefits Quarterly,Employee Benefits Journal, and Healthplan: TheMagazine of Trends, Insights and Best Practices.

Salisbury has served on the Secretary of Labor’sERISA Advisory Council, the Presidentially ap-pointed PBGC Advisory Committee, as a consultantto numerous government agencies and privateorganizations, and on committees of many profes-sional organizations. He is a Fellow of the NationalAcademy of Human Resources. He has written andlectured extensively on economic security topics. Heholds a B.A. degree in finance from the Universityof Washington and an M.B.A. in public policy andadministration from the Maxwell School at Syra-cuse University.

James P. SmithJames P. SmithJames P. SmithJames P. SmithJames P. Smith is a senior economist in the Humanand Material Resource Policy Department at theRAND Corporation in Santa Monica, CA. He alsoserved as director of the Labor and PopulationStudies Program at RAND from 1974 to 1994.Smith has served as principal investigator on anumber of projects, including an analysis of theeffects of economic development on labor markets; astudy of black-white wages and employment; recenttrends in women’s wages and labor force growth;the economic impacts of marital dissolution; life-cycle decisionmaking regarding consumption andsavings; racial income differences; the measure-ment and causes of income inequality of individualsand families; and asset accumulation of matureadults. In addition, Smith has participated inprojects studying the evaluation of economic loss inwrongful death cases, the labor supply effects ofincome maintenance programs, and the market forcollege graduates. Smith received a B.S. in econom-ics from Fordham University and a Ph.D. ineconomics from the University of Chicago.

C. Eugene SteuerleC. Eugene SteuerleC. Eugene SteuerleC. Eugene SteuerleC. Eugene Steuerle is a senior fellow at the UrbanInstitute. Earlier in his career, he was deputyassistant secretary for tax analysis at the U.S.Department of the Treasury. He directed theTreasury’s 1990 report, Financing Health andLong-Term Care, and was the original organizerand economic staff coordinator of the Project forFundamental Tax Reform, which led to the TaxReform Act of 1986. In addition, Steuerle hasserved as the director of finance and taxationprojects and resident fellow at the AmericanEnterprise Institute and federal executive fellow atthe Brookings Institution. He has published manybooks, reports, and articles, as well as providing

xiii

congressional testimony. His most recent books areThe Tax Decade and Retooling Social Security forthe 21st Century. Steuerle received his Ph.D. inpublic finance from the University of Wisconsin.

Lawrence H. ThompsonLawrence H. ThompsonLawrence H. ThompsonLawrence H. ThompsonLawrence H. Thompson serves as principal deputycommissioner of the Social Security Administration.Previously, he was assistant comptroller general incharge of the human resources division of the U.S.General Accounting Office. He served as the chiefeconomist of GAO from 1983 to the beginning of1988, when he assumed his current position. Hereceived his M.B.A. from the Wharton School of theUniversity of Pennsylvania and his Ph.D. ineconomics from the University of Michigan.

Jack L. VanDerheiJack L. VanDerheiJack L. VanDerheiJack L. VanDerheiJack L. VanDerhei is associate professor at TempleUniversity and also an EBRI Fellow and EBRIFellows Program Research Director. Prior to this,hw was a faculty member at the Wharton School ofthe University of Pennsylvania for eight years. Heis a member of the Pension Research Council andhas served as a consultant for the Pension BenefitGuaranty Corporation, U.S. Department of Labor,

and the International Foundation of EmployeeBenefit Plans. VanDerhei holds a B.B.A. andM.B.A. from the University of Wisconsin and anM.A. and a Ph.D. from the Wharton School.

Paul Yakoboski Paul Yakoboski Paul Yakoboski Paul Yakoboski Paul Yakoboski is a Research Associate with EBRI,specializing in pension and retirement incomesecurity issues. His current research focuses onretirement plan sponsorship and participationtrends, 401(k) plans and worker decisionmaking,lump-sum distributions and benefit preservation,the future retirement income security of today’sworkers, and the education of workers in partici-pant-directed plans. Yakoboski is also Chair of theResearch Committee for the American SavingsEducation Council and he is a member of theNational Academy of Social Insurance. He has aB.S. in economics from Virginia Polytechnic Insti-tute and State University and a M.A. and Ph.D. ineconomics from the University of Rochester.Yakoboski worked in the Human Resources Divi-sion of the U.S. General Accounting Office beforejoining EBRI in 1991.

1

Executive Summary

■ IntroductionThe report of the 1994–1996 Advisory Council onSocial Security settled one thing: we stand at acrossroads in deciding the future of the nation’sretirement system.

But we are ill-prepared to choose which pathto take. The options proposed by different factionsof the divided council—large-scale governmentinvestment in the stock market, a new mandatorysavings plan for individuals, or partial“privatization” of the current system—would leadus in profoundly different directions. Yet weunderstand only vaguely the impact each wouldhave on our retirement security, social relation-ships, and economy.

On December 4, 1996, the Employee BenefitResearch Institute (EBRI) hosted a policy forum toexplore how we can better assess these and otherSocial Security reform alternatives. The session,partly an exploration of the intricacies of math-ematical modeling and partly a preview of policydebates to come, gave some of the nation’s leadingauthorities on Social Security a first look atSSASIM2, a state-of-the-art computer model thatEBRI is developing to sort through the web ofdemographic, economic, psychological, social, andpolitical factors that lie behind various proposals.

Their reaction could best be described ascautious enthusiasm: EBRI won praise for its boldeffort but was reminded about the considerableuncertainties and guesswork involved in anyproject that is so complicated. Beyond that, manyparticipants said, some of the biggest challengeswill arise after the model is built, when the implica-tions of its results will have to be interpreted for apublic that has barely started to focus on the toughchoices that lie ahead.

Still, it was widely agreed, the modelingexercise is an important start in what is sure to bea long journey. “When you begin to look at these

issues, you discover right away that a great deal ofsystematic analysis is necessary to even describethem,” noted William Beeman, vice president anddirector of economic studies for the Committee forEconomic Development. “And even more sophisti-cated analysis is necessary to prescribe policies todeal with them.”

■ The Context: Pitfalls andPossibilities

In the most fundamental sense, the issues sur-rounding Social Security today are as old as theprogram itself. “Here we are again debating therole of social insurance in a market economy, tryingto define the appropriate balance between indi-vidual and collective responsibility, [and] what isfair and efficient at individual, family, and aggre-gate society levels,” noted Robert Friedland,director of the National Academy on Aging.

These are difficult questions, but recentreform proposals pose analytical challenges farmore daunting than the relatively modest, incre-mental changes that have been considered in thepast. Instead of seeking a single change—anadjustment in the payroll tax rate or a rise in thenormal retirement age, for instance—many currentproposals would modify diverse parts of the systemsimultaneously and add entirely new elements,such as self-directed individual accounts, to SocialSecurity’s traditional, defined benefit structure. Tocomplicate matters further, reform advocates saySocial Security should be judged in part by criteriathat were hardly considered in the past, includingwhether it provides beneficiaries a favorable “rateof return” on their tax payments and whether itfosters economic growth by encouraging savings.

“The complexity of the proposals undoubtedlyoutstrips our ability to carefully analyze them atthis point, and that’s the genesis of a model of this

ExecutiveSummary

Assessing Social Security ReformAlternativesby Christopher R. Conte


2

sort,” said Marilyn Moon, a Social Security trusteeand senior fellow at the Urban Institute.

Against this backdrop, EBRI President DallasL. Salisbury argued that we can’t continue to relyexclusively, as we have in the past, on eitherqualitative analysis or the work of the “extremelysmall” staff in the office of the actuary of the SocialSecurity Administration. Noting that advocates ofvarious reform alternatives employ tremendouslydifferent assumptions and methodologies, hedescribed the Institute’s modeling exercise as “aneffort to create an analytic capability that is able todo quantitative, apples-to-apples comparisonsacross proposals.”

While joining others in welcoming the under-taking, Moon added a note of caution, pointing outthat there are pitfalls as well as possibilities inmodeling. It’s often difficult to predict, she noted,whether trends from the past will continue into thefuture; can we assume, for instance, that womenwill continue to outlive men as they have in thepast, even though recent experience suggests thismay no longer be true? Moreover, she noted, somepolicies are particularly susceptible to modificationafter they are adopted, rendering earlier assump-tions moot. If people are given ownership of theirown Social Security accounts, she argued, theyinevitably will seek to use those funds for activitiesother than retirement; what confidence can wehave, then, about predictions concerning how manyassets they will carry into the later years undersuch a reform? In other cases, totally unexpectedfactors can neutralize the impact of variables weknow and understand.

“This is an interesting and important exercisebut is an exercise,” Moon said. “We all know, thoseof us who engage in this, that we’re going to bewrong. We just don’t know how we’re going to bewrong.”

Other participants pointed out that modelscan be helpful even if they can’t tell us with cer-tainty what will happen under different circum-stances. Beeman, for instance, argued that theyhelp force us to think more systematically. “There isa great deal of very loose thinking” about SocialSecurity right now, he said, pointing out, forinstance, that some reform advocates completelyignore the potentially sizable transition costs theirproposals would entail.

Martin Holmer, who is president of Policy

Simulation Group, EBRI’s contractor in developingthe Social Security simulation model, suggestedthat one of the greatest values of models can be toshow us not what the future will bring but ratherwhere the greatest risk of unanticipated or un-wanted outcomes lies.

Holmer gave conference participants a “lookunder the hood” at how the EBRI model works. Itstarts with the same demographic and economicassumptions used by the Social Security Adminis-tration but adds some variables, such as return onequity, that haven’t traditionally been consideredrelevant. It also analyzes some issues, such as thelikely return on individual investment accounts,that are beyond the scope of the current program.

Most significantly, the EBRI model employs“stochastic,” or “Monte Carlo,” techniques, whichenable analysts to develop not a single projection—or, in the case of the Social Security Administra-tion, “low,” “high” and “intermediate” projections—but rather a whole range of possible outcomes.Using probability theory, the model then shows thechances of various outcomes actually occurring.This is an improvement over traditional methods,most participants agreed, because it shows the oddsthat things will turn out worse, or better, thanprojected. With that information, policymakers candesign reforms that “hedge” against the biggestrisks, Holmer said.

■ The Mortality DebateOne example of the uncertainty surrounding themodeling process is the assumption we must makeconcerning the mortality rate. The Social SecurityAdministration assumes that mortality will declineat a far slower rate in the future than it did duringthe 20th century. But this view is controversial,noted Jack VanDerhei, associate professor atTemple University and research director for EBRI’sFellows Program. While the intermediate projec-tion by the Social Security actuaries assumes thatthe number of people over age 85 will rise to14.6 million in 2050 from 3.3 million in 1994, theCensus Bureau puts that figure considerablyhigher, at 18.2 million. Running the Census Bureaufigures through SSASIM2 suggests that SocialSecurity’s actuarial deficit will rise to 7.05 percentof taxable payroll in 2070, compared with the5.52 percent projected by the Social Security

3

Executive Summary

Administration.James Smith, senior economist at the RAND

Corp., said even the Census Bureau figures under-estimate the probable decline in mortality. He saida reasonable forecast would be that the number ofpeople over age 85 could total 21 million by 2050and that some “reputable scientists” believe it couldreach as high as 50 million, more than triple theSocial Security Administration’s estimate. More-over, Smith said official estimates fail to account forthe fact that higher-income people tend to livelonger than low-income people, thus adding dispro-portionately to benefit costs. That factor alone coulddrive costs 10 percent above official projections, heargued.

Stephen Goss, deputy chief actuary for theSocial Security Administration, defended theagency’s mortality assumptions. Not long ago, hesaid, a technical panel found that if Social Securitywere to change its mortality assumptions, it wouldhave to adjust them for people of all ages, not justfor old people. But a lower death rate among youngpeople would mean there will be more workers thanexpected paying payroll taxes, thus negating muchof the adverse financial consequences of largerpopulation of senior citizens. Moreover, according toGoss, the same panel found that the governmenthad underestimated future fertility rates, neutraliz-ing the effect of changed mortality figures entirely.

To hedge against the possibility of larger-than-anticipated gains in life expectancy, Holmer noted,the government of Switzerland has been exploringthe idea of indexing benefits to age expectancy fordifferent cohorts of the population. A less precisevariation of this strategy, of course, would be simplyto increase the normal age of retirement for SocialSecurity purposes. Robert Myers, who spent23 years as chief actuary for the Social SecurityCommission, recommended just that. Raising theretirement age to 70 by the year 2037 shouldn’teven be considered a cut in benefits, he argued,since by then a person who retires at that age willhave the same life expectancy as someone whoretires at 65 today.

■ Predicting Rates of ReturnIf Social Security actuaries have trouble projectinglong-term mortality trends, which they havestudied since the program’s inception in 1935, how

much harder would it be to estimate what returnsthe system could expect by allowing equityinvestment?

All three factions of the advisory councilassume that the historic advantage of equities overbonds will continue in the future (in recent decades,the yield on equities has averaged 7 percent morethan inflation, while Treasury bonds have paid only2.3 percent more). But Robert Shapiro, founder andvice president of the Progressive Policy Institute,warned that projecting future investment returns,for markets as a whole and especially for individualinvestors compared with other individual investors,will be very hazardous. “I question our ability tosensibly model behavior of financial asset and debtmarkets over time,” he said.

For individual accounts, in particular, howindividuals allocate their assets among differenttypes of assets will be a crucial variable. On thispoint, at least, EBRI has a growing base of informa-tion on which to build the model. Salisbury notedthat EBRI’s own defined contribution project nowhas data on how one million individual investorsallocate assets in their retirement accounts, andthat database eventually will grow to 10 million.

When it comes to asset allocation, the stakes,both for model-designers and investors, are veryhigh. VanDerhei produced figures from the simula-tion showing that a “life cycle” approach to invest-ing, in which a person puts his or her investmentsinitially in equities and then gradually switches tobonds over time will produce substantially betterresults than maintaining a consistent allocationover an entire lifetime. But even in the second case,where the hypothetical worker puts 40 percent ofhis assets in equities, the pay-off would be higherthan what would occur if the current Social Secu-rity structure were maintained and taxes wereraised to cover rising benefit costs or if benefitswere cut to fit the existing tax structure.

Significantly, the stochastic model alsoshowed that returns under the privatizationscheme would vary far more than under either ofthe “nonstructural” reforms. For critics ofprivatization, that was a crucial observation. “Howabout all the people who aren’t going to get averagereturns, and may even lose?” asked Robert Ball,former Commissioner of Social Security underPresidents Kennedy, Johnson, and Nixon and acritic of privatization. “I think we need to look at


4

the range of possible outcomes from these invest-ments—particularly for low-income people—andworry about whether they’re tolerable.”

Girard Miller, president and chief executiveofficer of ICMA Retirement Corp., echoed thatconcern and added another: Will the Social Security“safety net” have to be extended to people whohappen to retire during a bear market and hencehave to cash in or annuitize their savings whentheir value is low? The possibility of such “cohort-specific” market losses concerns a number of babyboomers, in particular. Some analysts believe atleast part of the run-up in stock prices in recentyears is driven by the baby boom generation’sgrowing demand for equity investments to helpfinance retirement. If so, some worry that theirretirement, and the resulting sell-off of their stockportfolios, could bring a long bear market.

“It’s been one of my constant worries thatwhen I come to retire and sell my 401(k) plan, I’mnot sure who I’m going to be selling it to,” saidWilliam Cheney, chief economist for John HancockMutual Life Insurance Co.

■ Macroeconomic EffectsPrivatization advocates believe Cheney’s fearswon’t be realized. If workers are encouraged, orcompelled, to save more, net investment would rise.If so, demand for the baby boomers’ portfolios couldbe strong despite the relatively smaller number ofbuyers. And, more importantly, increased savingswould lead to higher productivity and fastereconomic growth, benefiting society as a whole,they argue.

That appears to be the premise behind areform plan presented by the National TaxpayersUnion. It would gradually phase out the existingSocial Security system and replace it with a“national thrift plan.” Taxpayers would be requiredto set aside an amount equal to 5 percent of theirwages in “personal thrift accounts.” The govern-ment would match contributions of low-wageworkers, and guarantee a household income equalto the poverty level for all Americans over age 62.

While the plan would require sacrifices fromcurrent retirees by expanding taxation of benefitsand reducing cost-of-living increases, Howe pre-sented simulations that suggest its overall eco-nomic impact could be quite positive. Net nationalsavings would quintuple to 5.9 percent of Gross

Domestic Product by 2065, raising productivity andaverage wages by as much as 26 percent, he said.And he suggested that such gains would enableworkers at all income levels to receive much greaterbenefits than under the current system—allwithout new government debt, new taxes, orunrelated reductions in government spending.

“The savings-productivity-real wages link isabsolutely essential,” argued Neil Howe, a consult-ant who presented the taxpayer group’s plan. “It’sthis link that keeps Social Security reform fromturning into something close to a zero-sum game.”

Eugene Steuerle, a senior fellow at the UrbanInstitute, was skeptical about the economic-growthargument. Switching Social Security reserves out ofgovernment securities and into the stock marketcould unsettle the bond market and drive upinterest rates, especially if other governments wereto follow the federal government’s lead, he said.More fundamentally, the retirement of the babyboom generation will result in a drop in “humancapital” equal to a 20 percent or 25 percent rise inunemployment. Steuerle argued that it’s hard toimagine any level of capital investment that wouldproduce enough economic growth to boost incomeshigh enough to offset that dampening effect on theeconomy.

“I don’t think you can build enough steel millsto solve this problem,” he said.

Even the assumption that the reform propos-als would increase savings needs to be tested,Steuerle added. If the government creates incen-tives or requires more savings through SocialSecurity, he suggested, people might simply reducetheir savings outside the system by a comparableamount. As a result, reform proposals aimed atincreasing savings could wind up having “very littleeffect.”

While voicing doubts that Social Securityreforms will have much impact on economic growth,Steuerle joined most forum participants in urgingEBRI to press ahead in its study of the issue. TheRAND Corporation’s James Smith, for instance,argued that making no change in the program couldhave economic consequences just as significant asvarious reform proposals. According to Smith, thetax rate to pay for Social Security, Medicare, andother age-related transfers could climb to40 percent during the next century if current trendscontinue. That would lead to tremendous pressuresto cut other categories of government spending and

5

Executive Summary

could even choke off economic growth altogether.“We have to be talking about encouraging

savings and growth,” Smith said. He proposedimposing a progressive consumption tax.

Robert Myers, however, dissented from thegeneral sentiment that reform proposals should bejudged partly by what effect they would have on theoverall economy. “Social Security does not have thepurpose of solving all national problems,” he said,adding that economic growth, in particular, is “notSocial Security’s responsibility.” The real purpose ofthe retirement system, he said, is to ensure allretirees a basic “floor of protection.”

The importance of that floor was underscoredby Robert Friedland, who noted that61 percent of today’s elderly—72 percent of thoseover age 75—derive at least one-half of theirincome from Social Security. Similarly, JamesSmith noted that one-half of all retirees over age 70have financial assets totaling $10,500 or less. Forthose aged 51–61, the situation isn’t much better;the median level of financial assets for that group is$17,300.

Smith also argued that the current SocialSecurity system isn’t as progressive as it appears.Conventional figures on life expectancy suggestthat the typical minimum wage recipient willreceive lifetime benefits equal to a 4 percent rate ofreturn on his Social Security tax payments, com-pared with a 1.5 percent return for somebodywhose wages equal the maximum subject to SocialSecurity taxes. But if the figures are adjusted toreflect the shorter life expectancy for low-incomepeople, the minimum wage recipient will realize arate of return of only 3.1 percent, just slightlyahead of the 2.4 percent return for a wealthierperson collecting the maximum in benefits,Smith said.

■ The Perception GapWhile trying to untangle all of these analyticalquestions might seem to be challenging enough,several forum participants suggested that EBRI’sjob won’t be finished even after its model is built.

“Ultimately, what we want to do is try to findways to understand the implications of variousoptions and then to be able to explain them topeople in ways that are meaningful,” Moon said.“There’s a translation that going to need to be madehere.”

She argued that many Americans don’t fullyunderstand concepts like risk, the trade-off betweenrisk and return, and second-order effects, all ofwhich are crucial to the analysis of policy alterna-tives. “My caution today would be that we not be toocharmed with the elegance of the models and weopt for utilitarian approaches whenever possible.”

Susan Dentzer, chief economics correspondentfor U.S. News and World Report, underscoredMoon’s warning. “There is a very large potentialperception gap between many of us who work onthese very abstruse levels and the way these thingsare likely to be perceived by the public,” she said.

In particular, Dentzer questioned whethermany Americans understand financial matterssufficiently to invest their own Social Securityfunds. She also suggested that the “linchpin” ofprivatization, the assumption that an “appropriate”annuities market will develop and be capable notonly of having reliable returns but of securinginvestments once people retire, is far from assured.“Just hearing from financial market participantsthat it will take care of itself, frankly, does notincrease my confidence that this is a salable plan,”she said.

Even some analytically straightforward SocialSecurity reforms won’t be easy to sell, Dentzeradded. Raising the retirement age, for instance, willbe a “very, very difficult political propositionindeed,” she said. And she warned that there is agreater potential for “perceived social injustice fordifferent classes of individuals” in various reformproposals than experts recognize.

A number of forum participants urged EBRIto use its model not just to assess reform alterna-tives but also to develop a better understanding ofthe current Social Security system. Supporters ofthe existing system, for instance, voiced concernsthat the public misunderstands the system’scondition, believing it is in worse shape than itreally is. Myers dismissed as “myth” the belief thatSocial Security will go bankrupt sometime around2030; in fact, he said, the system’s low-cost estimateshows it will stay “in great shape” for the agency’sentire 75-year forecasting period. And members ofthe advisory council’s most traditionally mindedfaction point out that, even under the agency’sintermediate assumptions, Social Security stillcould cover three-fourths of its benefit costs afterthe year 2030 without any changes in policy.

Nevertheless, “no one has confidence, it


6

seems, that [Social Security] will be there,”Friedland noted. “I’m not sure that they reallybelieve it….Most people have not thought muchabout Social Security.…What little they knowseems to what they’ve heard, and what they’veheard comes from television and the newspaperand what they’ve heard is that the program is goingbroke.”

Gary Burtless, a senior fellow at theBrookings Institution, agreed that the press sharesblame for the public’s gloom. Seven years ago, hesaid, he co-authored the book, Can America Affordto Grow Old? Even though he and his colleaguesconcluded that the answer was yes, most pressaccounts and book reviews said the answer was no.“I think that’s a congenital problem among journal-ists, who may reason they will never see their namein print if they write a story containing either goodnews or at least ambiguous news,” Burtless said.

In any event, the public is eager to be edu-cated and is less susceptible to partisan efforts toexploit fears about Social Security or other entitle-ments than it has been in the past, according topolitical analyst and commentator Charles Cook.“They understand that we have some very seriousproblems with our entitlement programs,” he said.In a review of the 1996 elections, Cook argued thatthe Democrats succeeded for a time in chargingthat Republicans would cut too deeply into Medi-care and other entitlement programs, but the issue“evaporated” toward the end of the campaign. Inthe final analysis, he said, Democratic claims didn’tchange the outcome of any congressional race in thecountry. And, he said, people will require proofbefore they allow themselves to be frightened bysuch claims in the future.

Cook said he didn’t believe the politicalclimate was particularly favorable for entitlementreform, though. He predicted that many Republi-cans, feeling burned by the Democrats’ charges thatthey would wreck Medicare, probably will sit backand force President Clinton and his party to takethe lead on the issue. Many congressional Demo-crats, meanwhile, don’t feel a great sense of ur-gency about the matter.

When forced to act, many lawmakers willjump at easy solutions like making technicaladjustments in the Consumer Price Index thatwould slow cost-of-living increases in Social Secu-rity benefits, rather than enacting more fundamen-tal reforms, Cook predicted. Doing something that

appears painless, or that pins the blame on some-body else “would be very, very, very appealing tomost political people,” he said.

■ Reform AlternativesParticipants in the EBRI forum weren’t offeringmany painless solutions, though. During the finalphase of their discussion, they examined specificreform proposals, in the process demonstrating thatwhile policy prescriptions can be significantlyinfluenced by the results of economic modeling,they also are shaped significantly by views andjudgments that transcend science.

Some described the issue as essentially amoral one. Myers, who argued forcefully that SocialSecurity isn’t “broke” and shouldn’t be “thrownaway,” recommended raising the normal retire-ment age for Social Security to 70 by the year 2037and increasing the payroll tax by 1.2 percentagepoints for both employers and employees between2015 and 2035. He suggested normal economicgrowth should increase incomes enough that peopleeasily could afford such a tax hike. “But even ifthey didn’t rise, do Americans have to be sounaltruistic” as to oppose such an increase, heasked. “Do they have to be so concerned that theyalways have rising incomes and have five cars inevery garage and three television sets in everyroom?”

Many others based their reform plans on ablend of economic analysis, value judgments, andrealpolitik. Robert Ball, for instance, who was partof the advisory council group that recommendeddirect government investment in the stock market,seemed to defend the proposal more on politicalthan financial grounds.

“We have to do something about the percep-tion of younger people that they’re not getting agood deal under Social Security,” Ball said. “For thesecurity of the system, the people who are going tovote in the future need to understand and supportit and believe they are being treated fairly.”

Most of the projected Social Security shortfall,Ball argued, could be covered in “quite traditionalways” without changing the system’s basic struc-ture. His “maintain benefits” group on the advisorycouncil recommended, as first steps, extendingSocial Security coverage to state and local govern-ment employees, tighter taxation of Social Securitybenefits, modifying Social Security cost-of-living to

7

Executive Summary

would reach retirement with insufficient savings.That’s especially important, Thompson argued,because risk has increased in the private sector asmany employers have come to favor defined contri-bution plans over defined benefit plans. “If theprivate sector is getting riskier, we should be verycautious about transferring a whole lot more riskonto individual workers by significantly shrinkingthe public system,” he said.

While Ball and Thompson were concernedabout maintaining a strong role for the governmentin the retirement security realm, Ann Combs, aprincipal with William M. Mercer, Inc., and amember of the advisory council, made it clear thata desire for less government underlies privatizationproposals. Combs was part of the advisory councilfaction that developed the so-called “personalsavings account” proposal. It would allow workerseffectively to put 5 percent of their payroll earningsinto the personal accounts, which they would befree to manage and wouldn’t have to annuitizewhen they retire. The remaining 7.4 percentagepoints of the payroll tax would be used to pay a flatbenefit to all retirees, initially $410 per month. Inaddition, a tax equal to 1.52 percent of payrollwould be imposed to help finance the transitionfrom the current system to the new one.

Combs said the plan would guarantee retireesan income equal to two-thirds of the poverty leveland would eliminate some of the complexity of thecurrent benefit formula while creating a “verydirect link” between taxes paid and benefitsreceived. The personal savings accounts, mean-while, would lead to increased financial literacyand possibly would encourage more saving, shesaid.

Advocates of this approach contend that theindividual account plan would leave too muchcontrol in the hands of government. The “maintainbenefits” proposal of Ball and others would be evenworse in this regard, they argue. Although Ball saidthat government investment should be entirelypassive in its approach to equities investment,Combs questioned whether politicians could resistthe opportunity to let social objectives shape publicinvestment strategy. “In the end, it’s politicianswho make the decisions and can rewrite the rules,”she argued, “and I believe the temptation for socialinvesting or targeted investing.…will be too great.”

And while Ball and Thompson warned againstimposing more financial risk on retirees, Combs

reflect changes in the Consumer Price Index, andother changes.

But Ball acknowledged that all of thesechanges together won’t get the system completelypast the financial challenges posed by the retire-ment of the baby boom generation. Ball’s group saidthe payroll tax would have to be increased around2050. It also recommended taking a serious look atthe idea of investing up to 40 percent of SocialSecurity tax collections in the stock market.

Ball said the government should do theinvesting, rather than individuals, so as to main-tain support for Social Security as a social insur-ance program. Allowing people to set up individualaccounts in Social Security would sew the “seeds ofdestruction” for the system as a whole, he arguedbecause “average and above-average investors,”able to earn higher returns on their self-directedaccounts than on the traditional part of the SocialSecurity program, would demand that more andmore funds be switched from the regular benefitprogram into individualized accounts.

Lawrence Thompson, principle deputycommissioner of the Social Security Administration,presented the case for so-called “individualizedaccounts” with a different combination of economicthinking and practical politics. His starting pointwas that some combination of higher taxes andreduced benefits is inevitable. “If people are goingto live longer, either they’re going to have lowerbenefits in retirement, they’re going to put moreaway each year while they’re working, or they’regoing to have to work longer,” he said.

The individualized accounts proposal wasdeveloped by Edward Gramlich, dean of the schoolof public policy at the University of Michigan andchairman of the advisory council. It essentiallywould “split the difference” between benefit cutsand tax increases, Thompson said, by trimmingbenefits to a level that could be supported by thecurrent payroll tax and requiring individuals to payan additional 1.6 percent of payroll into individualaccounts. While individuals would have somecontrol over how funds in their accounts would beinvested, the choices would be limited, the moneywould be managed centrally, and all benefits wouldhave to be paid out in the form of indexedannuities.

Thompson said the public wants greaterindividualization. But he said limiting investmentchoices would help minimize the risk that workers


8

argued that the new financial market risk onworkers and retirees would be no greater than the“political risk” that they currently face—namelythat the government will change the rules of thegame by raising taxes or changing benefits.

■ What’s Next?Given the sharp divisions among experts, theuncertainty about what impact various reformproposals would have, and the fact that SocialSecurity’s projected insolvency is still some yearsoff, EBRI should have plenty of time to completedevelopment of its Social Security model. In themeantime, there’s a good chance we’ll see moreexperimentation and that may produce useful datathe Institute can plug into its equations.

Stanford Ross, a senior partner in the lawfirm of Arnold and Porter and a former SocialSecurity commissioner, proposed a cautious,pragmatic approach to reform that illustrates howwe might evolve gradually toward a new SocialSecurity arrangement.

The program should be brought into financialbalance, Ross said, partly to calm the “sky is fallingrhetoric” about the system. He also agreed thatthere should be a “personal account element”

because younger people are “less accepting ofgovernment paternalism.” But rather than settingout to solve the problem all at once, he said, weshould first adopt many of the incremental changesthat analysts like Ball have proposed and then letpeople voluntarily set aside additional funds inindividual accounts.

“One advantage of a voluntary approach is itwould give you valuable experience about how thepeople who would be affected really feel aboutputting away more for their retirement as opposedto consumption or other purposes,” he argued. “Youwould get valuable information.”

Ross’s recommendations, less comprehensivethan other proposals, would leave many questionsabout the future of Social Security unanswered.But in a sense, they summed up the current stateof policy analysis and policy making—includingboth our lack of information and the need forgreater understanding between the public andexperts. Without these, consensus may remainelusive.

“Any changes are going to have to be broadlybipartisan and based on a great deal of publiceducation,” Ross said. “Different people makedifferent calls on the economics and the politics andhow they think people react to things.”

9

Executive Summary

1A Framework for Analyzing andComparing Social Security Policiesby Kelly Olsen, Jack VanDerhei, and Dallas L. Salisbury

■ OASDI’s Finance IssuesUnder current law, the Social Security program willmeet the retirement of the baby boom generation1

in 2008, when the first boomers reach eligibility forearly retirement benefits at age 62. This retirementwave will only exacerbate pre-existing demographicpressures, primarily the result of our aging society,maturing social insurance systems, and lower birthrates in cohorts succeeding the baby boom genera-tion. In 1983, policymakers anticipated this long-range demographic strain by increasing the normalretirement age (NRA) and by raising Social Secu-rity taxes. Recalculating tax rates, policymakersaveraged the combined Old-Age, Survivors andDisability Insurance (OASDI) program’s cost as apercentage of taxable payroll over a 75-yearprojection period, resulting in a tax rate higherthan needed to fund short-term obligations.

In effect, this method of calculating theOASDI portion of FICA added a partial advancefunding structure to the Social Security systemthat went beyond the historical practice of simplymaintaining a contingency reserve. Due to the factthat, since 1983, FICA taxes have been higher thanneeded to meet current benefit payments, “surplus”Social Security revenue has been accumulated.

This revenue has been converted into Special-IssueTreasury bonds and credited to the Old-Age andSurvivors Insurance (OASI) and Disability Insur-ance (DI) trust funds, which are maintained by theTreasury Department. By the end of 1996, theOASDI trust funds had accumulated approximately$566 billion in assets, an amount anticipated topeak at about $1.3 trillion (in 1996 dollars) by theyear 2015.2

Theoretically, the Social Security trust fundsurplus will be drawn down as demographicpressures mount, helping younger workers pay forSocial Security retiree benefits and thereby keepingfuture FICA taxes lower than they would be if thesystem were maintained on a purely pay-as-you-go3

basis.4 Under intermediate assumptions, principalfrom the trust funds will begin to be used in 2019 tofinance the portion of Social Security benefitsobligations not funded by current FICA taxes. Inthe absence of reform, the 1996 Social SecurityTrustees’ Report estimates that the trust funds willbe depleted in 2029. At that time, FICA revenuesalone will be able to finance only about three-fourths of benefit obligations for the remainder ofthe 75-year projection period (2029 through 2070).

Few among the general American publicrealize that trust fund balances are dwindling by

1 Persons born between 1946 and 1964.

2 See Board of Trustees, 1996 Annual Report of theBoard of Trustees of the Federal Old-Age and Survi-vors Insurance and Disability Insurance Trust Funds(Washington, DC: Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability InsuranceTrust Funds, 1996). The trust funds held about13 months of Old-Age, Survivors and DisabilityInsurance (OASDI) benefits in reserve at the end of1996. Under intermediate assumptions, the trust fundis expected to peak with holdings of about 25 monthsworth of benefits by the end of 2005. Under the sameassumptions, the trust fund is projected to hold about13 months of benefits in reserve by 2015.

3 A pay-as-you-go system is one in which all FICAtaxes collected today are used to pay for all SocialSecurity benefits due today. That is, in a pay-as-you-go system, the only money used to pay current benefitsis money collected from current workers’ wages.

4 Note that the assumption that trust fund surpluseswill help future workers fund future benefits has neverbeen unanimously accepted. See Alicia Munnell andLynn Blais, “Do We Want Large Social SecuritySurpluses?” New England Economic Review (Septem-ber/October 1984): 5–21, and Robert Myers, “SocialSecurity and the Federal Budget: Some Mirages,Myths, and Solutions,” Journal of the AmericanSociety of CLU and ChFC (March 1989): 58–63.


10

legislative design, and, therefore, falling trustfund balances are not “news.” The real news aboutthe trust funds is that they were not expected todwindle as quickly as current projections predict.After passage of the Social Security AmendmentsAct of 1983, the 1983 Social Security Trustees’Report projected that the trust funds would hold 54percent of outlays in reserve by 2060 under thesecond set of intermediate assumptions (Board ofTrustees, 1983). In contrast, intermediate assump-tions used in the 1996 Social Security Trustees’Report project the OASDI trust fund balance to beexhausted by 2029.

Although legislative changes from 1983 to1995 and more optimistic demographic assump-tions had positive implications for the combinedOASDI trust fund balances, these were outweighedby other factors. The markedly more negativeprojections in the OASDI Trustees’ reports from1983 to 1996 are attributable to use of stricteractuarial methodology in calculating trust fundbalances, a change to more pessimistic disabilityand economic assumptions, and other changes.5 Anadditional contributing factor is that the periodprojected in the 1996 report includes 12 deficityears in which demographic pressures will bestrong (see following table).

Because, as a whole, the 1983 Trustees’assumptions are optimistic in retrospect, some areconcerned that 1996 Trustees’ projections areoptimistic as well and are therefore understatingthe OASDI long-range financial shortfall.6

In addition, critics of the current systemargue that the trust funds are already essentiallydepleted because their assets are borrowed by thefederal government (i.e., Congress), which usesthem to finance other government operations.When the OASDI program’s Treasury bonds mustbe redeemed in order to pay benefits, the only way

5 For example, the revised test to determine the trustfunds’ long-term financial condition became stricter in1992, and the methodology used to generate theeconomic assumptions was also changed. See MichaelAnzick, “1991 Social Security and Medicare AnnualReports Revise Insolvency Projections,” EmployeeBenefit Notes (August 1991): 1–8. The change inassumption generation means that “assumptions forthe future have been revised in a less optimisticdirection.” See Eugene Steuerle and Jon Bakija,Retooling Social Security for the 21st Century (Wash-ington, DC: Urban Institute Press, 1994).

6 In particular, program solvency is most sensitive tomortality and nativity assumptions, an area ofcontroversial debate even within the federal govern-ment. For example, the Census Bureau’s mid-rangeprojections predict 3.6 million more persons aged 85and over by 2050 than the OASDI Trustees’ mid-rangeassumptions. Some academics project that numberswill be even higher. See “U.S. Population Projections:2050 Ages 85 and Older,” National Institute on Agingand Census Bureau estimates, 1996.

Table 1.1Reasons for Changing Trust Fund Balances,

1983–1996

75 Year Actuarial Balance in 1983 OASDIa Trustees’ report = +0.0275 Year Actuarial Balance in 1996 OASDIa Trustees’ report = –2.17

(as a percent of taxable payroll) Change in Balance = –2.19

Reason for –2.19 Amount and DirectionChange in Actuarial Balance of Change

Legislation +0.10Valuation Period –0.55Economic Assumptions –0.79Demographic Assumptions +0.83Disability Assumptions –0.70Methods –0.93All Other –0.15

Source: Office of the Actuary, Social Security Administration, 1997.aOld-Age, Survivors and Disability Insurance.

for the government to repay its loans will be toborrow money from other sources, increase generaltaxes, or reduce other areas of government spend-ing. In any case, assuming that trust fund assetswill need to be tapped in order to pay benefits,general tax revenues are likely to rise. This leadsmany to speculate that the combined OASDI trustfund “isn’t really there” in the sense that the moneycurrent workers are paying today in excess ofcurrent benefit obligations is not going to help reinin overall tax rates tomorrow. Others argue thatoverall tax rates would be the same or even steeperin the future if the government had borrowedmoney from higher-interest lending sources in theprivate sector or had raised current income taxesinstead of borrowing OASDI trust fund reserves.Conversely, some speculate that the federal govern-ment would not have expended as many resourceshad the Social Security trust funds not beenavailable.

Whatever one’s perspective on the trust fundreserves’ efficacy in prefunding OASDI benefit

11

Executive Summary

obligations, projections show that the current FICAtax rate alone will be able to cover about 76 percentof projected program liabilities by 2029 (SocialSecurity Administration, 1996). Therefore, withoutraising additional Social Security revenues, benefi-ciaries in 2029 may receive only about three-fourths of what they are currently promised. Theprojected deficit over the 75-year actuarial periodfrom 1996 to 2070 is expected to be 2.17 percent oftaxable payroll under intermediate assumptions;7

that is, if payroll taxes were increased by thisamount in 1996, a 17.5 percent increase, thecombined OASDI program would be solvent until2070. Were the Congress to wait until 2022 toincrease taxes without cutting benefits, taxeswould have to rise 4.33 percent to 16.7 percent oftaxable payroll, about a 35 percent increase, inorder to keep the program solvent until 2070.8

As a result of this projected shortfall by 2029of roughly one-quarter of benefits promised,numerous and diverse reform proposals have beenpromulgated. Depending on their supporters’beliefs about the merit and viability of the currentsystem, these reforms range from fixing the SocialSecurity system in more traditional ways totransforming the existing system into a fundamen-tally different one (Advisory Council on SocialSecurity, 1997). The November 1996 issue of EBRINotes summarizes the main parameters of sevensuch reform packages (Olsen, 1996).

This paper provides a framework for evaluat-ing and comparing Social Security policies bydelineating 11 broad areas of consideration andhighlighting some of the relevant questions withinthese areas. This framework is not a comprehen-sive list of all considerations but is intended toprovide a feel for their complexity and to highlightsome of their most popularly recognized interactivepossibilities. As an introduction to EBRI’s SocialSecurity Reform Analysis Project, this chapter setsthe theoretical framework in which to place the

forthcoming technical results from the EBRI-SSASIM2 Policy Simulation Model. This modelwill produce stochastically generated quantitativedata regarding specific reforms.

■ Policy EvaluationSorting through the multitude of information aboutSocial Security proposals in order to fairly andcomprehensively compare and evaluate them canbe daunting. Those who do not abandon the effortaltogether are often tempted to make it moremanageable by focusing on only one or two aspectsof policy analysis, often spiraling into enormousdetail. Not only does this “blinders” approach risklosing the forest for the trees, but those who focuson only a few policy considerations inevitably fail toappreciate the interrelated nature of social pro-grams, economics, and other aspects of society. Achange in one aspect of a policy may not onlyresound throughout all its operations but may alsoresonate into other policy areas, affecting ostensi-bly unrelated aspects of the nation.

If nothing else, the complex evaluationframework delineated here is testimony to thepotential shortcomings of any proposals, regardlessof their supporters’ intentions or political align-ment, that offer professedly simple and sacrifice-free answers to Social Security’s finance issues. Inaddition, the framework lends itself to varyingdepths of analysis and comparison by presentingboth broad areas of consideration and by enumerat-ing specific questions within these areas. It shouldtherefore be a useful guide to both informedcitizens and policy analysts.

There are three broad areas to consider incomparing Social Security policies. The first isidentification of the policies’ underlying philoso-phies and assumptions that affect policy goals.Second, the nuts-and-bolts structure of each policymust be ascertained. Knowing the specific param-eters of a plan provides a starting point for predict-ing programmatic outcomes (i.e., the possible orlikely effects of the program for all those affected),the third broad area of consideration.

■ Underlying BenefitsDiffering ideas about the appropriate goals ofSocial Security policy emerge from varying underly-ing assumptions about political, economic, and

7 See Board of Trustees, 1996. It is interesting to note,however, that the 1996 Social Security AdvisoryCouncil assumes an adjustment of the consumer priceindex of 0.21 percent, which decreases the expectedshortfall from 2.17 percent of taxable payroll to1.86 percent. See Advisory Council on Social Security,Report of the 1994–1995 Advisory Council on SocialSecurity, Vol. I (Washington, DC: 1997).

8 Ibid.


12

intervention in redistributing resources affecteconomic growth? What is the inherent capacity ofthe private market to resolve or prevent socialproblems? Who is better at efficiently managingmoney, providing retirement security, mitigatingrisks, etc., the government or the private sector?

Approaching reform from different underlyingphilosophies means that advocates are at risk ofdebating on entirely different planes. While it isimpractical in time-constrained debate to focusheavily on identifying and evaluating advocates’underlying beliefs and attitudes, comprehensivepolicy analysis and comparison mandate recogni-tion of the impact they have on plan goals, on policydiscourse, and ultimately, on retirement incomepolicy and outcomes.

■ Program ParametersThe most fundamental structure in any retirementincome policy is basic plan design. After this isestablished, many aspects of how the programraises revenue and provides benefits follow. Inshort, the broad parameters of a policy are plandesign and program finance.

Basic Plan Design

Retirement income plans are designed as definedbenefit, defined contribution, or a combination ofboth. In a defined benefit plan, such as the currentSocial Security system, the goal is to guaranteethat, if workers participate, they receive benefits inan amount dictated by a predetermined formula.The government theoretically assumes completeresponsibility for fulfilling Social Security’s prom-ises by mandating contributions and by guarantee-ing benefit levels. To guarantee benefits, contribu-tions may have to increase over time.

In a defined contribution system, the plansponsor’s goal is to specify a method for determin-ing the cost of an individual account while provid-ing more participant choice. A consequence ofgreater choice is that the sponsor (in the case ofOASDI, the government) is absolved from responsi-bility for guaranteeing benefit levels at retirement.Although the government or employer may set therange of investment options available, participatingworkers shoulder responsibility for investmentgains and losses and therefore assume ultimateresponsibility for at least one part of their retire-

other human behaviors. In addition, policy advo-cates differ in their assumptions about demograph-ics and their views on desirable public policy goals,including beliefs about the appropriate delegationof responsibility between government and workersin providing retirement income security. The formerdifference is perhaps most pronounced in thecurrent debate, and is reflected in the fact thatreform packages range in fundamental design fromthose that would alter Social Security while main-taining the current defined benefit system (i.e., asocial insurance system that pools risk) to thosethat would supplement or replace the currentsystem with defined contribution (i.e., individual)accounts that return each dollar directly to the“family” that contributed the dollar.

Dissent about the appropriate responsibilitiesof government and workers in providing retirementincome can often be traced to differing basic beliefsabout human nature and government. This is anespecially pronounced difference that arose amongmembers of the 1994–1996 Social Security AdvisoryCouncil when it agreed to increase advance-fundingfor the OASDI program. In part because of differingbeliefs about who should maintain and manageadvance-funding reserves—individuals or govern-ment—the Advisory Council ended up factionalizedand proposed three different plans.

Related questions that bear relevance to thedesign of Social Security policy include: are peopleforward-looking and able to defer gratification inorder to plan for their own retirement; how muchpaternalism do people require in order to act intheir own best interests; how willing and able aremost Americans to inform themselves about andadjust to a new type of Social Security system? Theanswers to these questions affect one’s policy goalsbecause they are directly related to beliefs abouthow private individuals will respond to publicpolicy change. Similarly, opinions about govern-ment workers and political leaders influencepredictions of how the government will respond to anew policy.

Fundamental assumptions about the avail-ability and efficient control of economic resources isalso at the heart of much of the current SocialSecurity debate. How much can America afford tospend on Social Security while still tending to other“important” areas? In short, how much money isthere to “go around?” How does government

13

Executive Summary

mandated savings would reduce workers’ currentdisposable income, optional or mandated savingswould not transfer money from worker to govern-ment. Instead, optional or mandated savings rateswould be deferrals of a share of current wages intopersonal property vehicles for retirement savings.However defined contribution account assets wereinvested (so long as widespread default did notoccur), this approach would add revenue to theSocial Security program by bringing a greatershare of payroll into the system.

Whether financial solvency is maintained bylevying taxes, curtailing expenditures, issuing debt,investing program funds in the private market,and/or mandating savings, it is necessary to specifythe precise means of generating or saving programrevenue in order to identify all possible implica-tions. For example, if levying taxes or mandatingsavings is a means of generating program revenues,then the relevant questions are:• Who pays these taxes or contributes to the

individual savings accounts (e.g., the employer,employee, or others)?

• What would the rates be?• What is the source of the money (e.g., expendi-

tures on consumption goods, deductions frompayroll earnings, taxes on retirement incomeearnings, etc.)?

• What groups would bear larger tax or addi-tional savings contribution burdens, and whichgroups would be most affected by secondaryeffects from these changes?

If Social Security program expenditures are tobe reduced, questions arise as to which expendi-tures and which groups would be affected. In thecurrent debate, some expenditure cuts have beenaimed at current and future retirees’ benefits byadjusting the retirement benefit calculation.11

Changes in normal (NRA) and early (ERA) retire-ment ages, in the retirement earnings test, in the

ment security.9 Possible risks involved with beingultimately responsible for one’s own retirementincome are discussed in the section titled Adequacy(see p. 16).

Nevertheless, the government would still playa large part in promoting retirement security undera defined contribution-style Social Security system.At minimum, the government would likely beresponsible for providing or ensuring the availabilityof a vehicle for contribution collection and invest-ment as well as a means of benefit distributionthrough a system of individual accounts. In mostSocial Security reform proposals using definedcontribution accounts, the government is alsoresponsible for enforcing participation and, perhaps,for providing savings assistance to lower-incomeworkers.10

Program Finance

The finance parameter defines how a programmaintains financial solvency. Traditionally, govern-ment reaction to projected shortfalls in a federalprogram has taken three forms: raising taxes(“contributions”), lowering expenditures (“benefits”),or issuing debt. In addition to these traditionalmethods, private market investment is now apopular proposal for dealing with OASDI shortfalls.This approach attempts to raise program revenue bytaking advantage of stock market return ratesthrough investment of Social Security programrevenues and/or reserves in equities. It is assumedthat stock returns will continue to be higher thanbond returns.

Furthermore, some advocates of using definedcontribution accounts in the Social Security systemsuggest offering or mandating defined contributionsavings in addition to the current FICA tax pay-ments. This is another nontraditional approach tomaintaining program solvency in light of projectedshortfalls. While similar to a payroll tax in that

9 If one of the investment options were a guaranteedinvestment contract (GIC), or a similar guaranteedinvestment like indexed Treasury bonds, technically theissuer would ultimately shoulder some of the invest-ment risk, not the individual investor.

10 The National Thrift Plan proposed by the NationalTaxpayers’ Union Foundation is an example of adefined contribution-style reform that would requirethe government to assist low-income workers inmaking contributions to their personal accounts.

11 For example, the Maintenance of Benefits andIndividual Accounts Reform Plans offered by the 1996Social Security Advisory Council advocate usingworkers’ best 38 years of earnings, instead of thecurrently used best 35 to determine benefits. Averagingin these additional three years will, on average, reducelifetime benefits by 3 percent. See Kelly Olsen, “Keep-ing Track of Social Security Reform Proposals: ASummary,” EBRI Notes (November 1996): 1–8.


14

number of working years taken into accountin the Social Security benefit formula, in theconsumer price index (which is the basis for cost-of-living (COLA) adjustments),12 in spousal ben-efits,13 and efforts to de-link Social Security benefitlevels from the rise in real wages have all beenproposed.

If any portion of program funding is to comefrom the issuance of federal debt, the lenders mustbe identified in order to assess the implications forthe rest of the economy. For example, sale of debt tooverseas purchasers has different economic impli-cations than domestically purchased debt. Inaddition, the persons who would bear the burden ofpaying back additional debt need to be identified,along with the extent of their burden as determinedby interest rates, amortization period, and compet-ing demands on their resources.

If private market investment is included inthe reform, the next question is, who makesinvestment choices and who bears investmentrisks? If the government is to invest and assumeany inherent risks, rules about investment proce-dure and composition must be established to guidethe new reform and safeguard against possibleabuses. Specifically, the following considerationswould arise should the federal government becomea major stockholder in the private U.S. market:• In what percentage of the private market will

the government become a stakeholder?• If the government owns stock with voting

shares, will it vote and, if so, how?• Will the federal government voting its share

give the government “inappropriate” and/or“undesirable” control over the operation of theprivate market?

• If the government declines to vote its shares,then is it “problematic” that the remainingvoting shareholders will have disproportionateinfluence?

• How will guidelines be established for govern-ment investment performance reportingtimetables, and what will these guidelines be?

• How will the government choose investmentfunds, and will “social investing” occur,whereby the government chooses investmentfunds based on its charge to promote publicwell-being (e.g., not investing in a companyknown for Environmental Protection Actviolations or in a company that produces aproduct the Surgeon General’s Office deems asdetrimental to health)?

• Will investment be based on economicallytargeted investments (ETIs)? (Explained morefully on p. 26).

• Will the government simply try to index14 thestock market?

• How will market gains and losses be actuari-ally accounted for over time? Will the conserva-tive rules applied to private pensions by theInternal Revenue Service and the PensionBenefit Guaranty Corporation be followed?15

If a reform is adopted that allows individualsto make their own investments in the privatemarket, then, again, rules about investment

12 All three Advisory Council proposals assume thatthe Bureau of Labor Statistics (BLS) will adjust theconsumer price index (CPI), which is the basis onwhich Social Security benefits are indexed annually.In addition, the National Taxpayers UnionFoundation’s National Thrift Plan and Kerrey-Simpson’s Strengthening Social Security Act of 1995presume a CPI adjustment of –0.5 percentage points(Ibid.). The Boskin Commission’s Report to the SenateFinance Committee states that, “while the CPI is thebest measure currently available, it is not a true cost-of-living index (this has been recognized by the Bureauof Labor Statistics for many years). Despite manyimportant BLS updates and improvements in the CPI,changes in the CPI will overstate changes in the truecost of living for the next few years. The Commission’sbest estimate of the size of the upward bias lookingforward is 1.1 percentage points per year. The range ofplausible values is 0.8 to 1.6 percentage points per

year.” See Michael J. Boskin et al., Toward a MoreAccurate Measure of the Cost of Living: Final Reportto the Senate Finance Committee from the AdvisoryCommission to Study The Consumer Price Index(December 4, 1996).

13 Ibid.

14 Index funds seek to mirror the performance of thestock market by investing in every stock in an index.Broadly based index funds are used to ensure adequatediversification. One of the most common is theS&P 500, which tracks the stock performance of 500 ofthe largest domestic corporations. However, an indexcould be more inclusive and include smaller companies(e.g., the Russell 3000 index). Other questions wouldnecessarily arise during the implementation of thisstrategy. For example, should Social Security invest innondomestic equities or investments other than equitysecurities?

15

Executive Summary

procedure and composition must be established toguide the new program and to safeguard againstpossible abuses. In addition, if investments areheld by individuals, issues of property ownershipand control arise, such as:• Are contributions mandatory or optional?• Where can the individual keep his or her

individual account (e.g., can accounts be held inthe individual worker’s bank or brokerage firmof choice or will they be maintained by thegovernment)?

• Are individual accounts inheritable wealth?• Can workers access their individual account

funds for purposes other than retirement?• Should the federal government be able to

control investment options?• Similarly, will workers be required to hold

investments using a “life-cycle approach”whereby downside market risk is reduced asretirement age approaches by decreasingequity investment as age increases?

• Will workers be required to purchase or createlife annuities at the time they retire, or willthey be allowed to withdraw large lump-sumdistributions?

• If annuitization is optional, what is thegovernment’s obligation to persons who ex-haust their retirement funds prior to death?

Finally, a structure based on individualaccounts may create new roles for the federalgovernment. A defined contribution-style SocialSecurity reform plan must specify the government’srole in levying any new taxes, mediating invest-ment risks, helping low-income workers makecontributions, issuing or facilitating retirementannuities, and in meeting any other perceivedneeds arising from the new system’s definedcontribution component.

■ Comparing OutcomesAttempting to forecast and compare outcomes ofprograms that have not been implemented is themost difficult step in making fair and comprehen-sive comparisons, as determining program outcomeinvolves a myriad of sometimes very controversialassumptions. In addition, sometimes even ostensi-bly very small variations in expectations (e.g., one-quarter percentage point differences in annualgrowth, mortality, retirement, or interest rateassumptions) can, compounded over many years,result in vastly different policy outcomepredictions.

Moreover, outcome anticipation is difficultbecause the basic assumptions and fundamentalbeliefs that go into the creation of different pro-gram goals reemerge as points of difference. Forexample, if one believes the economy works betterthe freer it is of government intervention, thisbelief might cloud one’s prediction concerning areform that adds to government regulation. Simi-larly, if one views the economy as working best onlywith increased government intervention, one mighthave negative predictions concerning a reform thatreduces the government’s role.

Although people’s beliefs and assumptionsdiffer, and therefore so will their policy outcomepredictions, it is critical that they at least ask thesame comprehensive analysis and comparisonquestions. Asking the same questions establishes abasic starting point for debate. In addition, if only afew considerations are used, both informationrequired for identifying alternatives to currentpolicy and the comparison will be incomplete.

For the purposes of Social Security debate,there are 11 broad policy outcome considerations:(1) adequacy, (2) equity, (3) monetary costs,(4) other economic effects, (5) effects on the rest ofthe U.S. retirement system, (6) governmentaleffects, (7) administrative effects, (8) politicaleffects, (9) social effects (or, nonmonetary costconsiderations), (10) protection against uncertain-ties, and, finally, (11) the determination of the bestpolicy by weighing each of the aforementionedconsiderations. These outcomes must be identifiedin terms of three key aspects:• each outcome’s effect on all demographic

groups possibly affected,• potential short-term and long-term interactions

15 If the government were to invest directly in equities,the combined OASDI trust funds’ appearance of beingfunded over the short-term would vary because ofvariable returns from the private market. Currently,Special-Issue Treasury bonds provide fixed returns.However, the movement of Social Security definedbenefit assets into investments with more volatilereturns could subject the Social Security trust funds tothe same type of conservative rules that apply indetermining the funding status of private definedbenefit plans.


16

among outcomes, and• the rationale behind each outcome prediction.

The four points of consideration for outcomeprediction rationale are: assumptions used, validityof assumptions, sensitivity of outcomes to assump-tions, and robustness of these assumptions overtime (i.e., the likelihood that the predicted outcomewill persist). Clearly, sound outcome comparisonand analysis defies cursory examination and is amultifaceted process, as further detailed below.

Adequacy

Adequacy measures the degree to which policyobjectives are met in terms of benefit provision.16

In this context, adequacy determination is valueneutral. In examining a Social Security system,vertical adequacy considers whether beneficiarieswould receive the benefit levels indicated by plangoals. (Would actual benefit levels be as “high” asthe policy’s goals?) Horizontal adequacy askswhether all persons targeted to receive benefitswould actually receive them. (Would programcoverage be as “wide” as the goals suggest?) Forexample, would some not receive benefits becausethey do not know they are eligible or because ofadministrative errors? Finally, subjective adequacyconsiderations ask whether enough beneficiarieswould receive enough benefits (i.e., are policy X’soutcomes good public policy in terms of adequacy?).

A significant amount of discussion aboutadequacy has surrounded the current OASDIfinance debate, as adequacy, in terms of a nearpoverty-level income base or floor of protection, isone of the Social Security program’s two primarygoals.17 Because Social Security proposals usingdefined contribution accounts have been popular,potential risks involved in allowing individuals toassume more responsibility for their retirement

income security have raised concerns. Namely, thetwo primary risks are: (1) an individual’s likelihoodof using a risk-averse investment strategy, causinghis or her returns to be too low to yield an adequateretirement income base or floor and (2) theindividual’s risk of selecting equities or corporatebonds that provide a lower than expected invest-ment return experience.

To determine the likelihood that people willchoose risk-averse investment strategies, expectedinvesting behaviors on the part of individualworkers have become a paramount adequacyconcern. Questions about how to predict investmentbehaviors under a Social Security system thatinvolves defined contribution accounts include:• Which historic patterns of investment should

be considered (e.g., is the recent past moreilluminating about the future than the distantpast? Should we look at investment behavior inall retirement plans, including those with bothdefined benefit and defined contributioncomponents or just those with defined contribu-tion-only plans)?18

• Would these patterns be likely to continue if aSocial Security System with defined contribu-tion accounts were adopted (e.g., would peopleinvest more conservatively if they no longerhad a Social Security defined benefit as a“safety net”? Conversely, would people becomebetter at investment by necessity or because ofthe educational component of the new system?).

• Would the new system require a life-cycleinvestment approach (see below) to mitigatethe possibility of “age inappropriate” invest-ment choices?

Investment predictions tie into adequacyoutcomes by asking:• Are these predicted investment patterns likely

to result in the level of retirement security

16 Ibid.

17 The other primary goal of the OASDI program isequity. See Robert Myers, Social Security, Fourthedition (Philadelphia, PA: Pension Research Counciland the University of Pennsylvania Press, 1993).Adequacy can have many definitions. Many analystsdefine adequacy as replacement of the preretirementlifestyle during retirement. Social Security seeks toprovide this level of adequacy for very low-incomeretirees through a richer benefit formula than thatwhich applies to others. At maximum wage base

levels, adequacy is a life-sustaining benefit that is asmall percentage of preretirement income.

18 Some hypothesize that individual defined contribu-tion account investment behavior on the part ofworkers with defined contribution and defined benefitplans is more aggressive than that of workers whoexpect to rely solely on defined contribution plans fortheir employer-based retirement income. The rationaleis that the more dependent an individual is on aretirement plan, the less risk that individual is likelyto take with the assets of that plan.

17

Executive Summary

intended by plan advocates (vertical adequacy)?• If not, how could investment behaviors be

altered?• If investment behaviors are to be altered

through an educational effort, will there beenough resources to educate all groups andthereby ensure horizontal adequacy?

Concerns related to the risk of losing moneyin the equities market have led many advocates ofdefined contribution-style reforms to recommendthat equity investment be restricted to broadlydiversified funds comprised of stocks with histori-cally strong performance records. If choices arethus limited, the possibility of adverse investmentexperiences becomes less of a concern than it wouldbe if individuals were allowed to hold undiversifiedor highly speculative portfolios.

Some fear the risk of vertical inadequacy ifretirees were forced to withdraw from their definedcontribution accounts for living expenses or convertthese defined contribution assets into annuitiesduring a period of economic downturn. Mostfinancial planners suggest that, with age, individu-als should gradually convert investments inequities into investments with less volatile rates ofreturn. This “life-cycle approach,” which averts acrisis whereby retirement funds must be pulled outof a down market, could be adopted by SocialSecurity investors voluntarily or through a govern-ment mandate. Therefore, this particular risk ofvertical inadequacy depends on use of life-cycleinvestment strategy, not only on the performance ofthe market at any point in time. However, becausesome individuals may choose a life-cycle investmentapproach and others may not if it were optional,questions of horizontal equity emerge (discussed inthe next section).

Recall that adequacy, like every outcome

consideration, must be identified in terms of whatdemographic groups are most affected by program-matic adequacy lapses. In addition, it is importantto understand how adequacy predictions weregenerated as well as what effect the adequacy levelprediction may have on other considerations (e.g.,political or social considerations or the burdensplaced on other government programs).

Equity

Equity is the current OASDI program’s secondprimary goal (Meyers, 1993). Equity considerationsinvolve identifying policy outcome “winners” (thosewho will or do receive disproportionate benefits)and “losers” (those who will or do bear dispropor-tionate costs). Lifetime equity is often measured byascertaining replacement rates, internal rates ofreturn, and money’s worth ratios.19 Subjectiveequity considerations first include determination ofwhether programmatic distinctions made betweenbeneficiaries’ level of benefits (e.g., eligibilityrequirements and redistributive components) arefair. The following questions emerge:• Does the program reward certain behaviors,

e.g., continuous work force participation,increased savings behavior, etc., with greaterbenefits, and are these criteria fair?

• Is the program redistributive and should it be?If so, how much?

• How much should spousal benefits provide?• Are like beneficiaries treated alike, e.g., are

there “notches?”20 Are persons likely to receivedifferent benefits solely as a result of invest-ment “luck,” and is this an acceptable outcomeprovided that each beneficiary has equalinvestment opportunities?

A second equity consideration includesidentification of the distinctions made between

19 Replacement rates are the ratio of benefits payableat age 65 or the onset of disability to pretax earnings inthe prior year. Some analysts also define them asreplacement of average lifetime earnings. Internalrates of return are the rates of return at which thepresent discounted value of future benefits is equal tothe present discounted value of taxes paid. Money’sworth is the ratio measuring the present value of thebenefits a typical individual has received or is expectedto receive compared with the present value of thepayroll taxes and other contributions that he or she haspaid or is expected to pay, discounted at the actual past

and projected future rates of return on governmentbonds held by the Social Security trust funds. AdvisoryCouncil On Social Security, “Comparison of Plans,”Report of the 1994–96 Advisory Council on SocialSecurity, Vol. I (Washington, DC, 1996).

20 “Notches” refer to situations when younger workerswith the same earnings records get lower real benefitsthan older workers. For an explanation of the notchissue, see Eric Kingson and Edward Berkowitz, SocialSecurity: A Policy Primer (London, Auburn House:1993).


18

contributors’ levels of burden. Are theredisproportionate cost burdens imposed upon onegroup or age cohort? If so, are these distinctionsfair?

Another equity question takes into accountboth costs and benefits: are rates of return fairacross groups and/or age cohorts? Should employercontributions on behalf of workers be used incalculating individual rates of return? 21 In addi-tion, is it more important to consider whetherparticipants have equal opportunity for receivingfair rates of return or to consider only the level ofbenefits actually received? Finally, equity, likeevery outcome consideration, must be identified interms of which demographic groups are mostaffected by programmatic equity lapses. In addi-tion, it is important to understand how equitypredictions were generated and to consider whateffect this prediction could have on other outcomes.

Program-Specific Monetary CostConsiderations

Comparing reforms involves projecting short-termand long-term program costs as a percentage ofGross Domestic Product (GDP), taxable payroll,and real dollar values in the following areas:

• one-time start-up costs,• transition costs,• operating costs,• administrative costs,22 and• borrowing costs.

Assuming cost predictions are valid, thesecosts must not only be averaged to ascertainactuarial balance over a given period of time; theirdistribution over time must also be identified. For

example, does the actuarial balance dip below zeroat any time? If so, what are borrowing and taxrates predicted to be at that time? Is the averagecost low but the distribution unstable in thatsometimes costs are very low while at other timesthey are extraordinarily high? Finally, cost out-comes must be identified in terms of the demo-graphic profiles of the people most likely to beaffected in the immediate future and over time.

Cost estimates themselves are only credible ifthe assumptions and calculations used to derivethem are valid. For this reason, it is crucial for faircomparisons to ascertain assumptions and dataused, examine their validity, and assess the robust-ness of the results to changes in the assumptionvariables over time. It is also crucial to apply asensitivity analysis23 to determine cost estimates’sensitivity to changes in assumptions. One policymay have low costs only if certain variables remainconstant and high costs if one variable changes.Another policy might not have the potential ofreaching the low costs that the first policy iscapable of reaching but may instead provide a moresteady and predictable cost estimate because of itsability to maintain its cost rate within a range ofeconomic scenarios.

Finally, cost outcome predictions must beviewed in relation to other outcome predictions inboth the short and long term. For example, propo-nents argue that a higher cost Social Securitypolicy that mandates individual private investmentwould only necessitate reduced consumption untilthe standard of living gains generated from in-creased availability of investment capital arerealized, at which point consumption would risehigher than its current rate. On the other hand,some argue that the cost of a Social Securityprogram with defined contribution accounts wouldinitially decrease because of higher market returns,21 Individual assignment of employer contributions is

not always the case in employment-based retirementplans.

22 Administrative costs are a particular concern for asystem mandating individual savings accounts for allworkers. Workers with very low earnings (e.g., teenag-ers with summer jobs, seasonal workers) mightaccumulate less in their individual accounts than theseaccounts would cost to administer and maintain. Forthis reason, some have proposed that very low wageearners should be exempted from mandatory savingscontributions to individual accounts (see R.J. Myers,“Statement to the Subcommittee on Social Security ofthe Committee on Ways and Means, April 10, 1997”).

23 For example, there is a fair amount of disagreementwith respect to the “best” assumption to use for im-provement in future life expectancy. Sensitivityanalysis on this variable would consist of running costestimates under some baseline set of assumptions(perhaps those used in the Trustee’s report) and thenrerunning them assuming life expectancy actuallyincreases faster (slower) than the baseline. The newcost estimates will be higher (lower) for each reformproposal, but the important point is whether therelative and absolute rankings of the alternatives varyas the life expectancy assumption is modified.

19

Executive Summary

but that costs would eventually increase whileequity returns fall as the market adjusts to the newallocation of resources. (Such possible macroeco-nomic interactive effects will be explored morethoroughly in the next section.)

Extra-Programmatic Economic Considerations

Labor Force Participation—Pension systems,whether public and private, can encourage ordiscourage certain behaviors. In particular, theamount of benefit, the benefit accumulation rateprior to retirement age, the age at which benefitsbecome available, and how future benefits changewith continued labor force participation afterretirement age affect workers’ labor-leisure choicesand savings-consumption choices. Therefore, anintegral part of evaluating any retirement policy isthe identification and examination of its intendedand unintended effects on labor force behavior.

Many reform plans, as well as the changes inthe current Social Security system already legis-lated, are structured to provide incentives to delayretirement. The accompanying argument is oftenthat, although people today are living longer thanthey did in the 1930s when the Social Securityretirement age was established, the NRA remainsat age 65 for full benefits and at age 62 for earlyretirement.24 While a gradual, two-year increase inthe NRA was legislated in 1983, some advocate amore rapid increase to age 67 than that scheduledunder current law and/or extending the NRA evenfurther.25 Nonetheless, many of these proposals’most ardent supporters recognize that raising theage for early and/or full retirement benefits is notwithout some controversial implications. Evalua-tion of policies to raise the early and/or full retire-ment age(s) includes the following considerations:• Should public policy be concerned with when

people retire, or should retirement age beviewed as a personal matter? If policy ought toattempt to influence retirement age, how do

proposals affect the following issues:• How will a rise in the age at which Social

Security retirement benefits become availableaffect workers with physically demandingoccupations who will be less likely to be able towork in their occupations at later ages?

• If many laborers cannot work to the age ofbenefit eligibility, what is society’s obligation touse resources to support these persons (e.g., bypaying disability benefits), and to what extentwill these costs offset the economic benefits ofdelayed retirement for the rest of the popula-tion?

• How will the private sector respond to in-creased labor force participation after age 62?Specifically, will private employers be allowedto raise the age of eligibility for private pensionplans to correspond with the change in SocialSecurity policy? If allowed, will they?

• Will age discrimination impede increasedparticipation of older workers in the workforce?

• Will there be enough jobs for older workersboth numerically and in terms of appropriate-ness for the level of physical activity mostpersons over age 67 can perform?

• How will an increased number of older workersaffect rates of unemployment, job opportuni-ties, and career advancement for youngerworkers?

• If employers are expected to continue to employaged workers, can they afford to do so at thepay increase schedules currently common inbusinesses, or will pay schedules designed torise with seniority have to be curtailed and towhat extent?

Savings—Unlike the current system, many SocialSecurity reform plans are designed to increaseworkers’ saving incentives and thereby lift netnational saving. Many argue that an increase innet national saving will boost the capital available

24 In 1935, average life expectancy at age 65 wasabout 77 for men and 78 for women. It has sinceincreased to 80 for men and 84 for women (AdvisoryCouncil on Social Security, 1997).

25 Two plans put forth by the 1994-1996 SocialSecurity Advisory Council, the Individual Accountsand the Personal Security Account Plans, wouldincrease the NRA more rapidly than current law,

resulting in an NRA of 67 for persons turning age 62 in2011, as opposed to 2022 under current law. Thesereform plans would index the NRA to longevitythereafter, estimated to be about one month every twoyears. In addition, Kerrey-Simpson’s StrengtheningSocial Security Act of 1995 proposes an NRA increaseto age 70 by 2029. See Kelly Olsen, “Keeping Track ofSocial Security Reform Proposals: A Summary,” EBRINotes (November 1996): 1–8.


20

for investment and therefore spur economicgrowth.26 Increased voluntary savings incentives ora program of mandatory saving raises severalquestions:• To what extent can the government affect net

national saving, since an increase in one areaof saving could be offset by a decrease in otherareas of personal and governmental saving ifoverall consumption is not also reduced?

• How fast can the economy grow, and what isthe ability of a national retirement system toincrease growth rates?

• In a nation with many competing needs forresources, what is the appropriate priority levelof increasing national saving, i.e., how muchcan America afford to save?

• How would individual and employer contribu-tions to employment-based defined contributionplans be affected by higher Social Security taxrates and/or mandatory defined contributionsaving plans?

Market Interactions—In addition to providingincentives to promote certain worker behaviors, aSocial Security system exerts influence over thebehavior of both government bond and privateinvestment markets.27 As indicated, many SocialSecurity reforms have been specifically designed topromote saving so as to boost investment capitaland thereby spur economic growth. However, somepotential market effects may be unintended. SocialSecurity policy’s intended and unintended influenceon the behavior of private investment and govern-ment bond markets is an (often highly technical)area of increasing interest among economists.

Social Security (OASDI) affects privatemarkets because, with 141 million workers partici-pating as of 1995 (and with several plans proposedto increase that number),28 the program collectsand allocates large sums of capital. In 1995, incomeof the Old-Age and Survivors Insurance (OASI) andDisability Insurance (DI) programs was almost

$400 billion, which included a $59.7 billion surplusinvested in Special-Issue Treasury bonds (Board ofTrustees, 1996). By the end of 1996, OASDI pro-gram income was expected to have exceeded 5percent of the nation’s GDP, according to intermedi-ate assumptions (Board of Trustees, 1996). There-fore, increased incentives through the SocialSecurity program to save and/or to invest in privateequities could prompt reallocation of significantcapital and affect markets for this reason alone.Likewise, any changes in the amount of moneyhandled by the Social Security program, thedistribution of that money, or the investmentthereof could also affect the behaviors of privateand public markets.

The complexity with which changes in theOASDI program could affect markets is illustratedby consideration of the potential effects of the threeSocial Security Advisory Council proposals. Allthree proposals involve the reallocation of SocialSecurity assets from Special-Issue Treasury bondsto private-sector investments in equities in order totake advantage of the higher rates of returncurrently available from the private equity market(Combs, 1996). Treasury bonds are projected toyield annual rates of return that exceed inflation by2.3 percent, which appears low relative to equityreturn projections of about 7 percent over inflation(the same average as that over the past severaldecades) (Aaron, 1996; Social Security AdvisoryBoard, 1996). Advocates of investing Social Securityassets in equities expect the average return tocontinue to outperform rates on Special-IssueTreasury bonds.

An initial concern about investing SocialSecurity funds in the private market is the abilityof the private sector to absorb such large additionalresources. Hence, one consideration for a SocialSecurity system that would reallocate funding intoprivate markets from other parts of the governmentor society is how quickly this transition wouldoccur. Other concerns involve the expectation and

26 For example, the National Taxpayer UnionFoundation’s Thrift Savings Plan, the PersonalSecurity Plan and reforms advocated by Jose Pinera ofthe Cato Institute are designed to spur economicgrowth through increased private-sector investment.See Kelly Olsen, “Keeping Track of Social SecurityReform Proposals: A Summary,” EBRI Notes (Novem-ber 1996): 1–8.

27 The private investment market includes corporatebond and corporate equities.

28 For example, all three Advisory Council planswould mandate that new local and state governmentemployees be covered under the Social Security system.See Ann Combs, Social Security: Options for Reform(Washington, DC: William M. Mercer, Inc).

21

Executive Summary

money that the government borrows in thefuture?

• If interest rates on government borrowingsrise, will the interest on the federal debt riseand prompt an increase in income taxes?

• If income taxes rise as a result of a new SocialSecurity policy that reallocates revenue intothe private market, how much will the in-creased tax rates offset the benefits of the newpolicy?

Retirement Annuity Provisions29—A SocialSecurity reform with a defined contribution compo-nent could provide optional or mandatoryannuitization. Like concerns that the equitymarket would be unable to absorb the flow of assetsfrom a defined contribution-style Social Securitypolicy, there are concerns about how quickly andadequately the private annuity market would beable to adjust to increased demand. In addition,there is some debate about the impact on annuityprices of optional annuitization. Some believe thatincreased demand would modify the adverseselection inherent in a voluntary annuity marketand thus lower annuity prices, whereas otherspredict that the price of purchasing an annuitywould ultimately remain the same.30

Mandatory annuitization upon retirementraises the following equity considerations:• In the past, the employee’s gender was a factor

in determining the annual retirement benefitthat could be provided under a defined contri-bution plan. If a male and a female were thesame age and had exactly the same amountaccumulated under the plan, the male em-ployee would receive a higher annual pensionthan the female employee. This was becausethe female employee was expected to livelonger and, in anticipation of this, the sametotal amount was expected to be paid over a

likelihood of equity returns remaining as high (ornearly as high) as they have averaged over the pastseveral decades.

Depending on the reform plan, switchingSocial Security assets from Treasury bond invest-ments to equities could mean less demand forTreasury bonds initially because the TreasuryDepartment, a major purchaser, would no longer bebuying as many of them, if any (e.g., assuming abalanced budget by 2002). To attract investors, theTreasury might need to increase its interest rates ongovernment bonds.

If interest rates increase, some economicforecasters theorize that at least two forces couldlower returns on equities. First, higher demand forequities might mean that equity prices could be bidup, causing their long-range rates of return to fall.In addition, higher interest rates might hinderbusiness investment activities, which would belikely to cause the market value of equities to fallbecause of expectations of lower rates of return.Some argue that this effect could be mediated ifinternational markets are large enough. In fact, noone knows precisely how the markets would respondto a shift of Social Security assets from Treasurybills to equities, because the economic effects ofreallocating Social Security assets is a challengingarea that scholars have just begun to address.However, given a reallocation of Social Securityassets, the assumption that equity returns wouldremain at their current levels is not one to be takenwithout careful thought about market interactions.

Related considerations concerning the inter-connected nature of markets include the following:• If a new Social Security policy requires the

government to repay part or all of the money itborrowed from the Social Security trust fundsooner than under current law, would taxeshave to be raised sooner and more steeply?

• Similarly, if the federal government is prohib-ited from borrowing any additional revenuefrom the OASI and DI trust funds, can it beexpected to reduce its annual budget deficit inresponse by raising taxes or reducing spending?

• If the government does not reduce its annualspending, will it borrow money from the privatesector to replace the money it would otherwiseborrow from the Social Security program?

• If the government borrows more money from theprivate sector, will interest rates rise on the

29 For a detailed explanation of different types ofannuities, see Graves, 1994.

30 Much of the debate focuses on load factors, particu-larly the portions represented by the cost of adverseselection. For an excellent historical study of the costof adverse selection in the private annuity market, seeMark Warshawsky, “Private Annuity Markets in theUnited States: 1919–1984,” Journal of Risk andInsurance (September 1988): 518-528.


22

longer period of time. Because of the difference inlife expectancies, the actuarial value of thepension, in both cases, was considered to be thesame. In 1983, the Supreme Court ruled (inArizona Governing Committee v. Norris) thatlife annuities under an employment-baseddefined contribution plan must be provided ona uniform basis.31

• Unlike the current Social Security system,annuities in the private market are not in-dexed. Because women tend to live longer,inflation has more time to erode their purchas-ing power in retirement, and therefore lack ofindexation disproportionately affects women.

• While guaranteeing a spousal benefit if thebeneficiary is married, the current SocialSecurity system provides equal benefits toworkers with the same covered earningshistory regardless of marital status. The onlyway for a nonworking spouse to receive abenefit under a private market annuity is if ajoint spousal annuity is purchased. Unlike theSocial Security system, receiving a jointspousal annuity from the private marketmeans that, in exchange for joint spousalbenefits, the working spouse accepts a lowermonthly annuity payment.

• Would joint spousal annuities be mandatory forindividuals who qualify for Social Securitybenefits but whose spouses do not? If not,would this mean that nonworking spouses’retirement security would be left to the benevo-lence of their partners?

• If some nonworking spouses were not includedin their partners’ annuity contract, how muchwould society have to spend in terms of govern-ment programs to alleviate these spouses’potential poverty in old age? How much wouldthis expense offset the advantages of theprogram?

• Could annuities issued by the Social SecurityAdministration avert the above equity con-cerns? If so, how?

The foregoing questions highlight a differencein the fundamental objectives of the current Social

Security system and the current private annuitymarket system. In terms of providing guaranteedpayments for life, the private annuity marketemphasizes individual equity considerations thatseek to equate individual contributions to expectedindividual benefits on an actuarial basis, leavingindividuals with the responsibility of ensuring theadequacy of their own and their spouses’ retirementincome. In contrast, the Social Security systemcombines the goals of individual equity and socialadequacy so that benefits are based on need as wellas contributions (albeit indirectly) (Meyers, 1993).For example, if individual A paid twice the contri-butions as individual B and all other factors werethe same (e.g., age and gender), A would expect toreceive approximately twice the monthly benefits ina private annuity market system. Under thecurrent Social Security system, A would expect toreceive more than B (all things being equal);however, the benefit would be less than twice theamount B receives.

If annuities were not mandated and lump-sum distributions or any type of periodic paymentsfrom Social Security private accounts were permit-ted, individuals might desire to self-annuitize theirsavings in order to ensure a stream of income overthe course of their retirement. Self-annuitization isa strategy that an individual can use to ensure thathe or she does not outlive a particular amount ofprincipal. This is accomplished by dividing theaccount balance each year by his or her life expect-ancy at that point in time and limiting the annualconsumption to the amount determined by thecalculation. This step is repeated each year, and theannual amount will vary from year to year as aresult of investment income and changing lifeexpectancies. The requirements involved in suc-cessful self-annuitization raise concerns about:(1) the willingness and ability of retirees to performthese annual calculations and to practice fiscalrestraint, and (2) the fact that a certain percentagedrop in the markets would yield a proportionatedrop in annual consumption. As a result, theappropriate role of government or business pater-nalism is raised.

do not offer such annuities on a unisex basis, althoughlegislation that would require this has been proposed.Even though not required to do so, however, manyinsurers provide for unisex premiums.

31 It should be noted that employees can buy annuitiesfrom insurance companies on the open market (i.e.,apart from the qualified plan). At this time, insurers

23

Executive Summary

Effects on the Rest of the U.S. RetirementSystem

The result of the Social Security debate couldpotentially have great impact on the design of theemployment-based pension system. As mentionedearlier, an increase in the NRA and/or early retire-ment age would probably result in adjustments inemployers’ willingness to retain older employeesand the designation of retirement ages for employ-ment-based pension plans. (See the discussionunder Labor Force Participation and Savings onpages 18 and 19.)

In 1993, nearly 48 percent of employees inmedium and large private establishments werecovered by defined benefit pension plans usingbenefit formulas that were integrated with SocialSecurity provisions (U.S. Department of Labor,1995).32 Hence, if Social Security benefit provisionschange, the employers of approximately 7.7 millionworkers will most likely have to readjust theirretirement plan formulas.33 The total number ofemployers who would have to readjust is evenhigher, as the above figure does not include employ-ees of small private firms who may also be partici-pating in integrated pension plans. Readjustingbenefit formulas for Social Security changes wouldentail an administrative burden in addition to anyother potential burdens imposed on private pensionsponsors under a new Social Security policy.

Some other considerations with respect topotential changes in Social Security policy onemployment-based pension sponsors are thefollowing:

• Would tax incentives for employment-basedpensions—public and private—be reduced ifSocial Security costs put pressure on otherparts of the federal budget?

• Would employees demand that employment-based pension plans be more generous underpossible benefit cuts resulting from SocialSecurity reform?

• Since employees who retire early sometimesreceive bridge benefits from their employment-based pension plans until they become eligiblefor Social Security, would an increase in theNRA raise bridge costs and reduce bridgebenefits as a result?

• How much in resources can employment-basedpension sponsors be expected to allocate inadjusting to new Social Security policy in atime when they, too, will need to prepare andprovide for demographic pressures on their ownplans?

• If part or all of Social Security’s current definedbenefit system were reformed to includedefined contribution accounts, would workersfeel uneasy about not having as much of adefined benefit guarantee in retirement andtherefore pressure employers to expand em-ployment-based defined benefit plans in termsof benefits and sponsorship? Would employersbe encouraged to abandon employment-baseddefined contribution plans?

• If Social Security policy is changed to encour-age more delayed retirement, will employerhealth care costs rise as a result of an older

retirement payroll tax has been paid by the employeron these wages.

For a complete explanation of integration provi-sions, see Chapter 8 (for defined contribution plans)and Chapter 14 (for defined benefit plans) of Everett T.et al., Pension Planning, Seventh edition (Homewood,IL: Irwin, 1992). Also see Employee Benefit ResearchInstitute, Fundamentals of Employee Benefits, Fifthedition (Washington, DC: Employee Benefit ResearchInstitute, 1997) and James Schulz and ThomasLeavitt, Pension Integration: Concepts, Issues andProposals (Washington, DC: Employee Benefit Re-search Institute, 1983).

33 EBRI tabulation from the U.S. Department ofLabor, Bureau of Labor Statistics, Employee Benefitsin Medium and Large Private Establishments, 1993(Washington, DC: U.S. Government Printing Office,1995).

32 Integration with Social Security can be done inseveral ways, but the basic purpose of integration is toallow employers to take credit for the fact that they arefinancing one-half of the payroll tax assessed for theSocial Security retirement benefits for their employees.In certain defined contribution plans, employers areallowed to contribute a fixed percentage of compensa-tion for all parts up to a specified level of compensationand then a larger percentage for compensation inexcess of that amount (up to the 401(a)(17) limit). Thepermitted disparity between the two percentages iscontrolled by Internal Revenue Code sec. 401(l).

Integrating a defined benefit plan with SocialSecurity is a more complicated procedure; however, theemployer is allowed to indirectly increase the generos-ity of the benefit provisions for employees earning inexcess of the maximum taxable wage base ($62,700 in1996) in recognition of the fact that no Social Security


24

work force? If so, will this reduce the funds avail-able for employment-based pension plans?

Not only is there concern as to the extentemployers sponsoring pensions will adjust tochanges in Social Security policy, but there is alsodebate as to how workers participating in employ-ment-based plans will alter their behavior and howthese adjustments will affect their retirementsecurity.• If private investment accounts are incorporated

into Social Security, would workers, uponseeing large accumulations in their SocialSecurity accounts, be less likely to invest inemployment-based plans?

• If so, will this negatively impact retirementsecurity, or will the accumulations in SocialSecurity accounts be enough to sustain secureretirement?

• If the new Social Security program has adefined contribution component with educa-tional efforts, will this increase workers’awareness of the necessity and benefits ofsaving as well as the potential effects ofinflation and thereby increase worker partici-pation in employment-based pension plans?

• If the reformed Social Security system provideslower expected benefits than today, will work-ers realize the need to increase savings in theirdefined contribution employment-based plansto the extent permitted by the employer?

Ultimately, changes in the Social Securitysystem could impact all legs of the retirementincome stool,34 potentially changing its veryconstitution.

Governmental Effects

The Social Security OASDI program is not the onlygovernment program that promotes retirementsecurity. For example, Medicare Part A and Part Bassist in covering the costs of acute inpatient careand short-term rehabilitation as well as physicians’

visits and outpatient procedures; Medicaid pays forlong-term care services for the impoverishedelderly; the Supplemental Security Income (SSI)program and food stamps program assist poorelders in meeting basic living expenses; and theoffice of Housing and Urban Development (HUD)sponsors low-income housing programs for seniors.

Because the OASDI program is but one part(albeit a large one) of the entire U.S. system toprevent poverty in old age, the effects of the currentsystem or a new Social Security policy on thenature of the nation’s entire system of old-ageassistance policy must be considered.• What effects will Social Security reform have

on the Medicare program, which is facing moreimmediate insolvency projections and higherpredicted cost growth?

• To what extent do potential cuts in SocialSecurity benefits simply force cost shifting ontoother programs that target the elderly? If costshifting occurs, to what extent, if any, does theprogram stigmatize beneficiaries by eitherdirectly means-testing Social Security or bymaking more elders dependent on other means-tested programs?

• To what extent does the maintenance of currentlevels of Social Security benefits detract fromthe amount of money available to other govern-ment programs related to the elderly?

• To what extent does the redistributive generos-ity of the Social Security program discouragework and savings, thereby promoting depen-dence on government programs in retirement?

Just as the Social Security program is onlyone part of the federal government’s efforts tosecure retirement, these programs to benefit theelderly are but one part of the entire U.S. federalgovernment’s operations. Considerations related toother government functions include the effect ofSocial Security costs and benefits on:• other parts of the OASDI program that do not

directly target the aged (For example, willDisability Insurance (DI) be separated from therest of the system under a new reform? Willold-age and survivors benefits be separatedfrom survivors’ benefits for younger spousesand dependents? How will an increase ordecrease in the generosity of one OASDIprogram affect the resources and efficiency of

34 The retirement income stool has traditionally beendefined as having three legs: Social Security, indi-vidual savings, and private pension income. EBRIpublications, beginning in 1979, have suggested thatthere are more “pillars,” including wages from work,government assistance, survivor benefits, inheritances,long-term care insurance, etc.

25

Executive Summary

another?);• total public spending as a percentage of GDP;• federal deficit burdens;• resources available for other government

programs such as environmental protection,welfare to children, defense, public health,highways, education, and national security(opportunity costs from allocating resources tothe Social Security program that could havegone to other programs);

• amount of federal tax revenues collectible fromthe private sector; and

• confidence in the federal government and inthe Social Security program.

Administrative Considerations

Administrative costs are not the only administra-tive considerations that must be factored into policycomparison and evaluation. While related to cost,complexity of administration is a consideration inits own right, as increased complexity can lead todecreased program efficiency and thus to decreasedpolitical support. In order to assess administrativecomplexity, one must determine the point ofequilibrium whereby a program is complex enoughto meet the needs of a nation of different individu-als in various circumstances yet straightforwardenough to run efficiently. It is difficult to discernthe appropriate levels of complexity for the com-bined OASDI program, which covered 141 millionworkers and 43.4 million beneficiaries in 1995(Board of Trustees, 1996) and whose coverage iscontinually growing.

Administrative considerations include esti-mates of how often the policy will require regula-tory changes and of the policy’s flexibility adaptingto the nation’s ever-changing social, economic, andregulatory environments. Similarly, comparisonand evaluation require the prediction of howregulation will evolve. An initially simple policy canbecome a tangle of regulations over time underpolitical pressures. Finally, administrative consid-erations include identifying the office that wieldsadministrative power, evaluating the office’s pastperformance, and predicting its future behavior.

Political Considerations

Many political considerations involve the politicalfeasibility of passing and regulating a policy. Forexample, today’s Social Security debate includes

various opinions about how willing Congress wouldbe to mandate increased OASDI contributions ifsuch contributions were not “taxes” but “mandatorysavings contributions” instead. In addition, reserva-tions about regulatory feasibility are reflected inconcerns about administrative burdens and com-plexity.

Political risk issues seem to have dominatedpolitical considerations, however. Political risk isthe likelihood that a program or policy will lose itspolitical support or that policymakers will makechanges that prevent stated policy objectives frombeing realized. The political risk inherent in thecurrent system is evident in surveys that havefound high support levels for Social Security amongyounger workers but low levels of confidence thatthey would receive full benefits from the currentsystem.35 In addition, younger generations havelower expected rates of return on OASDI contribu-tions. As a result, some reform proponents believethat younger workers may provide less futuresupport for the program.

Another political risk is that policymakersmight wait so long to reform the current systemthat the changes needed for balance would beextreme, rather than limited, and cause politicalupheaval. A further risk relates to what mighthappen with a system of defined contributionaccounts were individuals to invest in the stockmarket. If a market drop were to occur (especially ifit happened right before a significant number ofpersons planned to retire), would the public de-mand restitution or a policy reversal?

A political risk concerning the use ofdefined contribution OASDI accounts to augment abase defined benefit program is the possibility ofdeclining political support for the defined benefitpart of such a “two-tier” system. Social Security’s

35 See Employee Benefit Research Institute, MathewGreenwald and Associates, and the American SavingsEducation Council, 1996 Retirement ConfidenceSurvey (Washington, DC: Employee Benefit ResearchInstitute, 1996); Employee Benefit Research Institute/The Gallup Organization, Inc., “Public Attitudes onSocial Security, 1995,” EBRI Report G-62 (Washing-ton, DC: Employee Benefit Research Institute, 1995);and Virginia Reno and Robert Friedland, “StrongSupport but Low Confidence: What Explains theContradiction?” in Eric Kingson and James Schulzeds., Social Security in the 21st Century (New York:Oxford University Press, 1997).


26

equity goal most favors middle and highincome participants, whereas its adequacy goalmost favors workers with lower incomes. Lowerincome workers tend to have a higher percentage oftheir preretirement incomes replaced by SocialSecurity, but those with higher incomes tend toreceive higher absolute benefits. A two-tieredsystem would make explicit which benefits origi-nate from the program’s equity goal (via the definedcontribution account, which would reflect contribu-tions and would thereby most favor middle andhigh income workers, who have the most to contrib-ute) and which benefits originate from the ad-equacy goal (via the defined benefit account, whichwould reflect entitlement and/or need, possiblymost favoring lower income workers). Would higherand middle income workers withdraw support fromthe adequacy (defined benefit) tier and press for theexpansion of the equity (defined contribution) tier,preventing a two-tiered program’s adequacy goalsfrom realization?

Additional potential risk involved in theprivate investment of Social Security funds, eitherby individuals or the government, are issues ofeconomically targeted investments (ETIs) andcorporate governance. Under a system of individualaccounts, unless workers were 100 percent free toinvest however they choose, the government mightestablish the range of “appropriate” investmentoptions available and any requirements for life-cycle investment. Among the considerations are:• Would criteria for selection of domestic and

international investment options for workersbe based on social considerations such as acompany or industry’s compliance with Envi-ronmental Protection Act standards or anation’s human rights practices?

• Would options be based on consistent returnshistory, such as on the basis of whether acompany’s stock were “blue-chip,” or would ETIoptions be mandated?

• If investment options were selected on thebasis of “blue-chip-type” standards or ETIs,would this unfairly discriminate against otherenterprises?

The government might exert more influenceon investment choices if it directly invested trustfunds in the private market itself, unless invest-ment is in index funds. Concerns are also raised

about real or imagined scandals between thegovernment and businesses that are vying forselection as a Social Security investment option oras part of the index in which Social Securityinvests. Such controversy could erode politicalsupport for the program and thereby expose definedcontribution-style Social Security plans to politicalrisk.

In addition, unless regulation of individualinvestment brokers is adequate, some fear market“scams” could emerge, which would underminesupport for a system using defined contributionaccounts. Others suggest that regulation of publicpension funds, such as the Federal Thrift Plan forfederal government employees, shows that thenecessary regulation and enforcement bodies are inplace to ensure that such quagmires can beaverted. However, this still leaves some concernedabout corporate governance issues should thegovernment own large blocks of stock in publiccorporations.

A final risk relates to defined contributionaccount balances. Would policymakers decide toallow preretirement access to funds under thepressure of other goals, as they have with indi-vidual retirement accounts (IRAs) and 401(k)s?And, if they did, what would be the retirementincome implications? Should people be allowedaccess to their accounts (through loans or earlydistributions) prior to retirement age for certaincircumstances or emergencies such as financingeducation, starting a business, paying medical bills,or avoiding bankruptcy? If yes, which circum-stances would be appropriate for early withdrawalor loans? Would early withdrawal or loans involve apenalty? If Social Security account balance holderswere unable to make their mortgage paymentsbecause of preretirement withdrawal or loanrestrictions, would they demand that the system bechanged? The history of access rules for privatepensions and IRAs suggests that this issue wouldbe regularly revisited.

Some oppose preretirement access to definedcontribution-style Social Security accounts, arguingit would undermine the system’s political stabilityby eroding its financial stability, sincepreretirement use of funds would affect the ad-equacy of postretirement benefits. Others supportpreretirement access in situations where a personwould otherwise need government support on the

27

Executive Summary

grounds that, ultimately, preretirement access forthese individuals would have no net effect ongovernment expenditures. If such persons couldwithdraw from their Social Security accounts intheir preretirement years, the government mayneed to support them during their old age; if suchpersons were denied preretirement access, thegovernment may have to support them until oldage. Those not opposed to preretirement accessargue that, either way, the government pays andSocial Security account accumulations are used.

Social Considerations

Social considerations involve the identification ofcosts and benefits that do not have a quantifiable,objective monetary value but are nonethelesscrucial to factor into the comparison and evaluationof Social Security policy. Social considerations thathave been identified include:• Are people informed about the financial

protection and risks involved in the program?(Some, for example, have raised the question ofwhether defined contribution-style programscould ever adequately inform younger workersof financial risks, since younger workers havelived in an era of relatively constant highreturns from equities, where risk existed butwas not realized on aggregate.)

• Does the program promote financial literacyamong participants?

• Does the program provide a sense of nationalcommunity and cooperation?

• Does the program contribute to peace of mindin regard to the adequacy of aged familymembers’ incomes and in one’s own retirementsecurity?

• Does the program help mediate the burdens ofyounger family members in caring for andfinancially supporting their aged?

• Is the program’s public discourse honest? (Forexample, are tax increases and benefit cutspresented straightforwardly to the Americanpublic, or are they hidden behind more politi-cally palatable rationales?

• Are Social Security costs being shifted to otherprograms without this being understood by theAmerican public? Does the program makeexcessive promises?

Because of different fundamental beliefsabout the appropriate nature of a Social Security

system, one may find all, some, or none of the abovesocial considerations appropriate to ask of SocialSecurity policy. What social considerations areconsidered germane depends significantly on thesubjective value framework and fundamentalbeliefs taken into account in determining appropri-ate OASDI policy goals.

Protection Against Uncertainties

Related to the foregoing social consideration ofpeace of mind, a Social Security program’s ability toprotect against uncertainties, or risk, means that atleast one part of individuals’ and households’retirement security cannot be eliminated by shocksin earnings or expenditure needs such as disability,unemployment, unforeseen longevity, or unexpectedinflationary growth (Boskin et al., 1996). Thecurrent system protects against inflation andunexpected longevity by indexing and annuitizingbenefits. Furthermore, the current system attemptsto protect all aged and/or dependent marriedpersons from the risk of losing support because of aspouse’s death by providing survivors’ benefitsunder the retirement program. Finally, the currentOASDI system mitigates the effects of shocks toearnings from disability by providing disabilityinsurance.

Trade-offs occur in a system that attempts tomitigate uncertainties. The primary trade-off is thelack of individual control that results from partici-pation in a program that pools risk. This meansthat some beneficiaries will receive less than theywould have if they had controlled their own moneyand not participated in the program; others, whoexperience misfortune such as disability, widow-hood, or unexpected longevity, will do better thanthey would have done on their own.

An underlying premise of many Social Secu-rity reform advocates is that the risk-poolingnature of social insurance programs actuallyincreases the likelihood of risk occurrence. Theyargue that if individuals know that risks leading topoverty in retirement-—such as unforeseen longev-ity, inflation, and disability—are guaranteed to beallayed should they occur, they will not do every-thing in their power to prepare for these risks.There is a wide range of opinions concerning whichand to what extent risks and uncertainties arecontrollable and the appropriate treatment ofpeople who take advantage of the risk-mediating


28

nature of the current system by insufficientpreparation. Hence, evaluation and comparisonrequire a conception of what level and type of riskprotections should be part of Social Security policyand consideration of whether a given policy actu-ally meets that standard.

Weighing Considerations to Determine theBest Social Security Policy

This final consideration recognizes the variation insubjective weights that different persons place onthe 10 other areas of consideration. For example,some may believe that social considerations are lessimportant than economic considerations, or vice-versa. Some place higher value on achievingadequacy in a Social Security program, whereasothers find equity a more compelling consideration.Some believe equity means redistribution, butothers believe it means proportionate returns.These value differentials determine how peopleweigh the importance of the above 10 consider-ations and thereby synthesize a large amount ofinformation into one decision about which SocialSecurity policy is “best.”

■ ConclusionJust as this evaluation and comparison frameworkbegan with considering the impact that policyadvocates’ subjective value bases and beliefs haveon plan goals, it ends with similar subjectiveconsiderations of policy evaluators’ weightings ofdifferent areas of outcome consideration. Hence,the framework emphasizes the role of fundamentalbeliefs and values in the Social Security policydebate, highlighting the risk that discourse amongdifferent parties may take place on entirely differ-ent philosophical planes. However, the SocialSecurity debate is not solely a subjective one; thisframework has presented some of the numerousand interrelated technical, economic, political, andadministrative questions that need to be addressedin the process of fair and complete analysis andcomparison.

The framework also shows that people withthe same value beliefs can arrive at different policyoutcome predictions, whereas people with the samepolicy outcome predictions can arrive at entirelydifferent valuations of Social Security policyoptions. The interplay between complex beliefs and

values is testimony to the potential shortcomings ofany proposed “easy answers” to the resolution ofSocial Security’s finance issues. As this frameworkshows, in fair and comprehensive comparison andanalysis, every Social Security policy reform, aswell as the current system, must be held account-able to a number of hard questions.

■ ReferencesAaron, Henry. “Behind the Privatization Rush.”

Washington Post, 22 July 1996, p. C1.Advisory Council on Social Security. Report of the

1994–1995 Advisory Council on Social Security.Vols. 1 and 2 (Washington, DC, 1997).

Board of Trustees. 1996 Annual Report of the Boardof Trustees of the Federal Old-Age and Survi-vors Insurance and Disability Insurance TrustFunds (Washington DC: Board of Trustees ofthe Federal Old-Age and Survivors Insuranceand Disability Insurance Trust Funds, 1996.

—————. 1983 Annual Report of the Board ofTrustees of the Federal Old-Age and SurvivorsInsurance and Disability Insurance TrustFunds (Washington, DC: Board of Trustees ofthe Federal Old-Age and Survivors and Disabil-ity Insurance Trust Funds, 1983).

Boskin et al. Toward a More Accurate Measure ofthe Cost of Living: Final Report to the SenateFinance Committee from the Advisory Commis-sion to Study the Consumer Price Index.December 4, 1996.

Combs, Ann. Social Security: Options for Reform.Washington, DC: William M. Mercer, Inc., 1996.

Graves, Edward E., ed. McGill’s Life Insurance.Bryn Mawr, PA: The American College, 1994.

Myers, Robert. Social Security. Fourth edition.Philadelphia, PA: Pension Research Counciland the University of Pennsylvania Press,1993.

Olsen, Kelly. “Keeping Track of Social SecurityReform Proposals: A Summary.” EBRI Notes no.11 (November 1996): 1–8.

Social Security Administration. “Actuarial Status ofthe Social Security and Medicare Trust Funds.”Social Security Bulletin (Summer 1996); 57–63.

U.S. Department of Labor. Bureau of Labor Statis-tics. Employee Benefits in Medium and LargePrivate Establishments, 1993. Washington, DC:U.S. Government Printing Office, 1995.

29

Executive Summary

2Economic Security: An Overview ofSocial Securityby David V. Bryce and Robert B. Friedland

■ IntroductionSocial Security is the nation’s largest and mostsuccessful public program. It has affected the waywe view retirement and enabled generations ofelderly to leave the labor force and remain finan-cially independent. Even though most people havelittle confidence in the future of the program, it hasand continues to be very popular (Friedland, 1994;Reno and Friedland, 1997). In 1995, an estimated141 million workers participated in the SocialSecurity system, and 43.4 million people, or17 percent of the population, received benefits fromit. Most recipients were retirees (62 percent), but9 percent were disabled workers, and 29 percentwere surviving dependents, including 2.3 millionchildren.

The passage of Social Security not onlychanged people’s lives, it changed our views aboutthe role of government and about retirement.However, this debate is not over. In fact, since thedays of the earliest pilgrims, we have been sortingout the boundaries between private and publicconcerns (Achenbaum, 1987). The public policydebates over Social Security were certainly consis-tent with this all-American struggle. Whether theprogram should be mandatory or voluntary; federalor state; its impact on the private sector; andindividual responsibility were all issues in thedebates. Our responses to these issues reflect oursocietal values. These questions must be continu-ally asked in light of changing demographics,economics, and expectations.

During the 104th Congress, the debate overindividual responsibility and government acceler-ated. However, Social Security was left out of thedebate. Social Security is not likely to remain “offthe table” forever. At the very least, the SocialSecurity Trustees will remind us that the combinedOld-Age, Survivors, and Disability Income (OASDI)Trust Funds will not be solvent beyond 2029.

Furthermore, Social Security is 22 percent of thefederal budget and will be growing. More impor-tantly, the 1995 Social Security Advisory Council’srecently released recommendations for ensuringsolvency to 2070 are sparking interest in reform.This debate is reminiscent of the debate surround-ing the initial passage of the Social Security Act in1935. This discussion reviews the past in light oftoday’s issues.

■ Enacting the Social Security Actof 1935

On August 14, 1935, approximately one year afterhe had charged the Committee on EconomicSecurity with the task of developing a governmentprogram capable of protecting Americans “againstthe hazards and vicissitudes of life,” PresidentFranklin D. Roosevelt signed the Social SecurityAct into law. The act’s foremost provisions estab-lished a compulsory old-age insurance program, anunemployment compensation program, and afederal-state matching fund program for assistanceto the aged, the blind, and the fatherless. Thelegislation was a watershed event in Americanhistory, enacting the first truly national publicbenefits program—and correspondingly establish-ing the federal government as a major player in thearea of social welfare—and providing a policyblueprint to which subsequent public benefitprograms would largely adhere.

The act’s passage is especially significantwhen considered in light of America’s historicalaversion to government initiatives. From its birth,the United States has generally maintainedallegiance to the Jeffersonian credo, “That govern-ment is best which governs least.” This is particu-larly true in the case of public benefit programs,which did not gain political legitimacy until theGreat Depression. Prior to that, political move


30

ments in the United States tended to concen-trate on reforming the democratic process—forinstance, expanding the right to vote or achievingdirect primaries—and rarely viewed the govern-ment as a potential contributor to material security.So deep was the aversion to public benefits thateven major labor unions opposed such programs,fearing that they would wean workers away fromunion loyalty and toward government dependence.Thus by the 1920s, when each country in WesternEurope had instituted national benefit programs,public benefits in the United States consistedprimarily of local poor relief.

America’s lack of interest in social welfarewas predicated on the popularly held belief that,through individual industriousness, all personswere capable of becoming self-reliant. In accordwith this view, the prospect of destitution wasconsidered a key ingredient in motivating individu-als to maximize their productivity. Many fearedthat if the government alleviated the prospect ofdestitution by instituting benefit programs, indi-vidual incentives to work would decrease, thusjeopardizing the economic and moral health of thenation. As Frederick L. Hoffman, an executive withPrudential Life Insurance in the early 1900s,stated in summarizing his opposition to publicretirement provisions, “pensions will undermine...the self-respecting character of our people ascitizens in a democracy where economic indepen-dence, achieved by individual effort, self-sacrifice,and self-denial, is, after all, the only aim and endworth while” (Lubove, 1968). In the late 19th andearly 20th centuries, it was generally presumedthat the poor were solely responsible for theirpredicament.

However, despite this philosophical convic-tion, a variety of arrangements were designed toalleviate economic suffering. Many of these wererelief programs operated by voluntary associationsor charities. Individual families were also a fre-quent source of support for aged relatives. In fact,many communities mandated that such support begiven, passing laws that punished family memberswho failed to provide for their elderly parents.

In addition to associations, charities, andfamilies, some individuals received support fromprivate benefit programs offered by their employer.These were generally designed to assist retirees orthe dependent family members of workers killed in

job-related accidents. In the early 20th century, andparticularly following the First World War, whenAmericans witnessed the “dole” rapidly spreadingthrough war-ravaged Europe, company benefitprograms were heralded for their ability to protectworkers from the hazards of an industrial economy,while simultaneously promoting the individualwork ethic. Pointing to the situation in Europe,advocates of company programs noted that privatesector welfare provisions were far less likely to fallprey to political pressures calling for liberal benefitexpansion. Yet, while scholars debate the motivesbehind private pensions, they agree that theseprograms were inadequate, despite their propo-nents’ claims.

In a study conducted in the 1930s, Murray W.Latimer estimated that no more than 14 percent ofthe labor force was covered by such a program(Latimer, 1932). Furthermore, many people whowere covered failed to receive benefits. According toLatimer, two-thirds of all noncontributory privatepension plans in 1932 contained disclaimers statingthat workers who had fulfilled their service andconduct requirements had no “right” to benefits.Consequently, in 1932, in a population of personsaged 65 and over totaling 6.5 million, less than2.5 percent received income from a private pension.

Although Americans were disinclined to lookon government as a source of material relief, somepublicly administered old-age pensions did existprior to the Depression. Most prominent was theCivil War Pension Program (CWPP), which by 1894accounted for 37 percent of the total federal budget.Funded by general revenues, the CWPP expandedrapidly in the years following the Civil War, oftenbeing used as a source of patronage to lure prospec-tive voters. This experience, combined with exces-sive corruption within the program, has led somehistorians to argue that the CWPP’s susceptibilityto political pressure for expansion played a role inconvincing the architects of the Social Security Actthat benefit programs funded through generalrevenues were neither financially viable norpolitically wise.

There was also a variety of state run benefitprograms. The majority were noncontributory old-age pension programs that empowered countyofficials to determine whether benefits would beprovided within their jurisdiction. Most frequently,they were not, and in 1928 only about 1,000 elderly

31

Executive Summary

persons were receiving state pensions. Moreover, anumber of state operated pensions were deemedunconstitutional. In 1928, the Supreme Courtoutlawed programs in Montana, Nevada, Pennsyl-vania, and Oregon. At the time, it was well estab-lished in the law that taxation could be imposedonly for “general” purposes. In the court’s view,public pension programs violated this principlebecause they raised taxes specifically to finance“individual” benefits.1

The Impact of the Great Depression

The Great Depression abruptly shattered the viewthat the only impediment to economic security washard work. Between 1929 and 1932, thecommonprice stock index plummeted, and realGross National Produce (GNP), which had in-creased by 22 percent between 1923 and 1929,dropped 30.4 percent. Five thousand banks, with

1 These rulings, and the precedent on which they werebased, greatly affected the drafting and content of theSocial Security Act. The Supreme Court in the 1930spossessed a decidedly negative opinion of governmentlegislation pertaining to social or economic issues.However, although fear that that act would be struckdown led to Roosevelt’s infamous attempt to “stack” thecourt, members of the Committee on Economic Securitywere aware of at least two rulings that augured wellfor public benefit programs. In the first, Frothinghamv. Mellon, the court established the constitutionality offederal grants in aid for the purposes of social provi-sion. In the second, Florida v. Mellon, the court uphelda tax offset credit device contained in the Revenue Actof 1926. Awareness that the court accepted thesepractices as constitutional influenced the SocialSecurity Act, particularly the components establishingold-age assistance and unemployment compensation.Both of these programs were financed through federalgrants in aid, and unemployment insurance containeda considerable tax offset device intended to encourageemployers to contribute to unemployment funds.However, neither of these rulings provided a window ofopportunity for old-age insurance, a program that wasto be the federal government’s exclusive domain andwas fully financed by payroll taxes. As a result, fearspersisted that the court would interpret the payroll taxas one intended to provide for “individual” rather than“general” welfare and thus void the act.

Several steps were taken to minimize the likeli-hood of this occurring. First, to de-emphasize theconnection between benefits and taxes, the act wascareful to detail these provisions under separate titles.Second, the act made no mention of directly relatingbenefits to contributory taxes and did not directlystipulate that the proceeds to the tax be placed in an

old-age reserve account. Finally, although proponentsbelieved that the key to popularizing the act wasequating contribution s with benefits, thus creating aperception that individuals were contributing to theirown retirement account, they were careful not topromote Social Security in this manner until after theSupreme Court had ruled on its constitutionality.

In 1937, the court handed down two rulingsestablishing the constitutionality of the Social SecurityAct. The first, Steward machine Co. v. Davis, vali-dated the unemployment insurance tax required ofemployers of eight or more employees. In brief, thecourt ruled that the tax was legitimate because it wasbeing used to promote the general welfare. Accordingto the opinion, written by Justice Cardozo, “It is toolate today for the argument to be heard with tolerancethat in a crisis so extreme the use of the moneys of thenation to relieve the unemployed and their dependentsis a use for any purpose narrower than the promotionof the general welfare.” In the second case, Helveringv. Davis, the court upheld the use of payroll taxes forthe provision of old-age benefits. This opinion, againread by Justice Cardozo, argued that the “generalwelfare” had been altered by industrialization in amanner that adversely affected the elderly. Accordingto Cardozo, “More and more our population is becom-ing urban and industrial instead of rural and agricul-tural. The evidence is impressive that among indus-trial workers the younger men and women are pre-ferred over the older. In time of retrenchment, the olderare commonly the first to go, and even if retained theirwages are likely to be lowered.” In the court’s opinion,the consequences of this trend were “national in areaand dimensions,” and hence a federal program of old-age benefits was ruled to be in the interest of thenation’s “general welfare.”

assets totaling $3.2 billion, became insolvent;90,000 businesses failed; and aggregate wages fellto 57.4 percent of their 1929 value. As unemploy-ment increased from 3 percent to over 25 percent,people became progressively aware that individualeffort provided no guarantee of material security,and they began increasingly to seek governmentassistance. This shift in public opinion culminatedin the presidential election of 1932, when HerbertHoover, vowing to continue “private efforts,” wasresoundingly defeated by Franklin D. Roosevelt, acandidate who pledged government initiative andinnovative solutions.

While the Depression laid bare the economicinstability of numerous social groups, it wasparticularly uncompromising to the elderly. Be-tween 1929 and 1932, 45 private pension planscovering over 100,000 workers ceased to exist.Private charities and relief agencies were swamped


32

beyond capacity and forced to turn awaythousands of needy seniors. According to economistand future Senator Paul H. Douglas, these circum-stances “convinced the majority of the Americanpeople that individuals could not themselvesprovide adequately for their old age and that somesort of greater security should be provided bysociety” (Achenbaum, 1987). In the early 1930s,this popular realization was increasingly channeledinto social movements that called for the implemen-tation of government pensions.

The Influence of the Townsend Movement

Although historians debate the relationship ofvarious social movements to the passage of theSocial Security Act, they generally agree that theTownsend Movement was integral to the inclusionof old-age insurance. Townsend, a doctor fromCalifornia, proposed providing each person over age60 with a monthly check for $200, stipulating onlythat recipients be retired and spend the entire sumwithin 30 days. Two hundred dollars a month was afantastic amount, especially considering thatworkers in 1932 took home average monthly checksof $95. Yet in Townsend’s view, which mimicked aprevalent economic theory of the time, the economycould only be rejuvenated through consumption.Providing this amount of monthly income wouldspur the economy by initiating consumption.

Townsend proposed to finance his planthrough a 2 percent transaction tax on business,and despite the unfavorable response he receivedfrom both the academic and business community,he rapidly accumulated a base of staunch support-ers. By 1935, his movement had 3.5 million dues-paying members, a club in every congressionaldistrict, and had submitted a petition to Congresssigned by 20 million people who urged the passageof a “Townsend Act.”

Legislators were clearly aware of publicsupport for the Townsend Plan. According to RobertDoughton (D-NC), who chaired the House Waysand Means Committee, the American system itselfwas in jeopardy: “The presence of insecurity onsuch a vast scale is a serious threat to our economicorder...the fact that several of these proposals haveattracted a widespread following implies a threat toour existing institutions which should not beregarded lightly.” Similarly, Congressman JamesMott (R-OR) asked of his fellow legislators, “Is this

body, the duly constituted representatives of thepeople, going to completely deny their petitions?”

Such rhetoric convinced President Rooseveltthat radical measures like the Townsend Plan—which his advisors informed him was financiallyimpossible and was in any case considered by thePresident to be incompatible with America’sproclivity for individual effort and reward due to itsreliance on a flat benefit structure—were notnecessarily political long shots. This concernedRoosevelt, who was willing to experiment withusing government to speed recovery but feared thatCongress might push him “in a direction far moreradical than any he had originally contemplated”(Leuchtenberg, 1963). Furthermore, Roosevelthimself was sensitive to a growing public desire forsome form of government benefits.

To avoid this, the president took the initiativein designing a government benefit program thatwould be sustainable, in accordance with hispolitical views, and capable of alleviating theeconomic insecurity gripping America. Thus onJune 8, 1934, he addressed the nation, stating thatthe foremost objective of recovery was “the securityof men, women, and children” (Witte, 1963).Roosevelt defined “security” as consisting of threefactors: a decent home to live in, the development ofthe nation’s natural resources in a fashion thatwould maximize employment opportunities, andprotection against “the hazards and vicissitudes oflife.” Regarding this latter concern, Rooseveltpromised to establish a committee to formulate aproposal for submission to Congress at the begin-ning of its next session. This became the Commit-tee on Economic Security (CES).

■ The Approach of the Committeeon Economic Security

The CES—headed by Secretary of Labor FrancesPerkins, who chose Professor Edwin Witte from theUniversity of Wisconsin to serve as staff director—was charged with developing a workable socialinsurance system. Any group then facing such achallenge could draw on two distinct approaches topublic benefits, both named after states that hadimplemented programs based on the respectivephilosophies. The Wisconsin Plan stressed preven-tion of economic insecurity, while the Ohio Planemphasized providing adequate benefits to those in

33

Executive Summary

need. Thus the Ohio Plan recommended that thegoal of benefit programs was to provide a “livingwage” to those in need, for instance, the unem-ployed or retirees. Conversely, benefits in theWisconsin Plan were not predicated on a directresponse to economic need but rather were in-tended to mitigate the impact of particular occur-rences such as the loss of income due to retirementor disability. Furthermore, the Wisconsin approachattempted to minimize interference with marketforces by linking retirement benefits to anindividual’s employment and wage history. Inessence, the Wisconsin Plan provided a floor ofprotection in the event of certain occurrences whileotherwise remaining as faithful as possible to theAmerican spirit of individual effort and reward.

Roosevelt clearly favored the Wisconsin Plan,believing the Ohio approach too closely paralleledthe public assistance programs proliferating inEurope and would ultimately lead to dependencyon “the dole.” When introducing to Congress thelegislation that would become Social Security,Roosevelt stated, “The lessons of history, confirmedby the evidence before me, show conclusively thatcontinued dependence upon relief induces a spiri-tual and moral disintegration fundamentallydestructive to the national fiber. To dole out reliefin this way is to administer a narcotic, a subtledestroyer of the human spirit. It is inimical to thedictates of sound policy. It is in violation of thetraditions of America” (Schiltz, 1970). ThusRoosevelt sought a benefit system that would bedistinct in both practice and public opinion from“relief” or “welfare,” based on his belief that the keyto avoiding such stigmas was the development of aprogram that attained its social legitimacy from theachievements of beneficiaries. As a result,Roosevelt overwhelmingly selected advocates of theWisconsin Plan to head the CES, preconditioningthe committee’s final report, which was presentedto the him in early January 1935.

Fulfilling Roosevelt’s wishes, the report wasquickly transformed into legislative language andsubmitted to Congress. Surprisingly, consideringthe magnitude of the bill, it met with relativelylittle congressional opposition and was not dramati-cally altered from its original form prior to passage.Legislators who did oppose the bill were those whogenerally rejected any form of government benefitprograms.

One of the most outspoken critics wasRep. Charles Eaton (R-NJ), who summarized hisopposition stating, “I think we stand today in thiscountry at the crossroads of a great decision whichtranscends all parties, all sections, and all inter-ests; and this decision is whether we are going tochoose American industry as the instrument for thesolution of these tremendous far reaching prob-lems, or whether we are going to resort to somemodified form of ‘Russianism’ and attempt to solvethese problems by government” (CongressionalRecord, 1935). In response to such hyperbolicappeals, Chairman Doughton expressed his beliefthat the act did not undermine private and indi-vidual work ethics. According to Doughton, thesewould be preserved by contributory financing, as“The worker’s right to benefits is conditioned uponhis previous employment, social insurance will donothing to break down the sacred American tradi-tion of self-reliance and initiative” (CongressionalRecord, 1935). Ultimately, the legislation wasopposed by only 37 congressional members.

The Social Security Act

In brief, the main provisions of the bill called for anunemployment compensation program, a compul-sory old-age insurance program, and a federal-statematching fund program for assistance to theindigent aged. Both unemployment compensationand old- age assistance were highly decentralized,with the former relying on a federal tax offsetdesigned to encourage states to establish andregulate their own unemployment compensationprograms and the latter being financed jointly bygeneral federal revenues set to match the fundseach state contributed to their own individuallyadministered program.

According to Edwin Witte’s memoirs, theseprograms received relatively little congressionalattention due to the legislators’ overarching inter-est in old-age pensions (Witte, 1963). Yet, twosubstantive changes were made that applied toboth provisions. The first was the decision of theWays and Means Committee to exclude certaingroups from coverage, most significantly agricul-tural workers and domestic servants. This wasdone at the behest of Southern legislators—whoheld a majority on the committee—and has beeninterpreted by most historians as being emblematicof the South’s fierce opposition to federal interfer


34

ence in the region’s handling of racial issues.The second was the elimination of a stipula-

tion in both provisions that state programs for old-age assistance and unemployment provide benefitlevels high enough to ensure “a reasonable subsis-tence compatible with decency and health.” Thisrequirement was vigorously attacked by Congress-man Harry Byrd, Sr. (D-WVA), who argued that itwould provide the federal government with extraor-dinary powers to arbitrarily influence local wagesand living standards. In the case of old-age assis-tance, the removal of this provision prevented thefederal government from establishing minimumbenefit levels, and resulted in Old-Age Assistance(OAA) programs that varied widely from state tostate in terms of benefit levels. For instance inMississippi, OAA paid only $3.92 per month in1938, while in California it paid $31.36.2

The major part of the congressional andpublic debate surrounding the Social Security Actinvolved provisions most germane to the act’s old-age insurance component. Considering the magni-tude of the legislation, deliberations regarding old-age pensions were surprisingly demur. Yet severalkey issues were raised, some of which continue toappear in contemporary debates on Social Security.Perhaps one of the most significant concerned anamendment offered by Sen. Bennet Champ Clark(D-MO). Clark proposed to exempt employersoperating private pension programs from the oldage insurance program. The administration wasopposed to such an exemption, fearing that it wouldresult in the public sector being saddled with theresponsibility of providing for an undue proportionof “high risks.” In this view, allowing employerswho offered a private plan to “opt out” wouldcompromise the program’s ability to spread costs,thus jeopardizing its financial stability. Yet despitethis opposition, Clark’s amendment passed theSenate by a vote 51–35 and was not dropped inconference until Roosevelt made it clear that hewould not sign the bill if it contained the exemptionclause.

In addition to the Clark Amendment, debatessurfaced regarding the use of contributory taxes,

the rate at which such taxation should occur, thesignificance of an old-age insurance trust fund andthe age at which benefits should be received, andwhether or not retirement should be a condition ofbenefit eligibility. These topics are the focus of thenext section.

■ The Debates Over ContributoryTaxes

President Roosevelt was a strong supporter ofcontributory taxes. Because this method of financ-ing tied benefits to work history, he believed it wascompatible with America’s political and economictraditions. Several years after the Social SecurityAct had been passed, Roosevelt was asked by areporter to summarize his reasons for insisting thatAmerica’s public benefits system be funded bycontributory taxes. He replied, “We put thosepayroll contributions there so as to give contribu-tors a legal, moral, and political right to collecttheir pensions and unemployment benefits. Withthose taxes in there, no damn politician can everscrap my social security program” (Achenbaum,1987). This statement points to one element ofpayroll taxes that attracted Roosevelt—institutinga system that linked contribution and return wouldconvince the public that they had a vested interestin the program.

However, during the formation of, and debateover, the Social Security Act, reliance on contribu-tory financing was not a given. In fact, there wassignificant opposition to full-scale contributoryfinancing, and the initial CES report itself calledfor government contributions to the old-age insur-ance program.

Contributory Taxes, the Initial CES Report, andTax Rate Adjustments

Just prior to the CES’s formation, Roosevelt statedthat the funds for public benefit programs “shouldbe raised by contribution rather than an increase ingeneral taxation.” In meetings with CES staff, thepresident frequently reiterated his conviction thatfull reliance on contributory financing was neces-sary to prevent the emergence of a public “dole.”Considering Roosevelt’s popularity and politicalstature, his opinions weighed heavily on the CESstaff. Thus it is somewhat surprising that the

2 In 1972, the Old Age Assistance program wasreplaced by Supplementary Security Income (SSI), aprogram that is entirely financed and administered bythe federal government.

35

Executive Summary

initial committee report recommended using fundsfrom general government revenues in 1965, theyear they predicted payroll taxes would no longercover old-age insurance program costs.

The CES called for general revenues in 1965primarily because it estimated that progressivelyless favorable worker-to-beneficiary ratios wouldeventually require either a reduction in benefits, anunscheduled increase in the payroll tax rate, or acontribution from general revenues. In the originalCES plan, benefits were to be paid at a rate set tomatch 5 percent contributions on payroll—2.5 percent each from employee and employers.However, the tax rate was not scheduled to increaseto a joint rate of 5 percent until 1957. The initialtax in 1937 was to be 0.5 percent on both employeeand employer, increasing by 0.5 percent every fiveyears until it reaching the maximum rate of2.5 percent in 1957. Thus any beneficiary who hadbegun paying payroll taxes prior to 1957 wouldreceive more than he or she had contributed.

This discrepancy would be minimized by ahigh ratio of contributors to eligible retirees duringthe program’s early years, but as this ratio evenedout, a financial imbalance would eventuallyemerge. In the estimation of the CES, this deficitwould begin in 1965, increasing to $1.4 billion in1980. The CES proposed that this deficit be met bygeneral revenues but acknowledged that there wasno way to ensure that it would not be accounted forby reducing benefits or increasing taxes.

On January 16, 1935, one day before he wascommitted to sending the CES report to Congress,Roosevelt became aware of the general revenueprovision. Calling Secretary Perkins to his office,Roosevelt ordered the financing provisions to beredrafted so that old-age insurance would beentirely financed by payroll taxes. To do this in oneday, tables in the report showing future govern-ment contributions were omitted and the relevantprose hastily altered. The report did concede thatthe plan’s financing would require revision in thefuture and a statement was added indicating that“there may be a valid objection to this plan in thatit involves too great a cost upon future generations”(Leff, 1983).

To rectify this situation, Roosevelt chargedthe CES and the U.S. Treasury Department withthe task of developing a fully self-financed old-ageinsurance program. The solution they devised

would be presented to the House Ways and MeansCommittee by Treasury Secretary HenryMorgenthau and hence become known as the“Morgenthau Amendment.” The revisionsMorgenthau proposed—and which were adopted—ensured that the old-age insurance program wouldbe sufficiently financed until at least 1980. Thiswas accomplished by raising tax rates, speeding upthe intervals at which these rates increased, andreducing pensions for early beneficiaries.

Under the new plan, the initial tax on bothemployers and employees was set at 1 percent,increasing by 0.5 percent every three years until amaximum of 3 percent was reached in 1949.However, these revisions not only ensured theprogram until 1980, they also generated a reservefund of $50 billion. The existence of this fund wouldbecome the source of considerable debate.

Opposition to Contributory Financing and theTrust Fund

In Roosevelt’s view, the problems that could belinked to a sizable old-age reserve fund wereinsignificant when contrasted with the drawbacksof using general revenues to finance public benefitprograms. In fact, some historians have suggestedthat Roosevelt was sympathetic to a large reservefund because of advantages he believed it wouldyield. Namely, they point to Roosevelt’s rarelyappreciated fiscal conservatism, noting that hesupported a balanced budget and only sacrificedthis principle in the face of the terrible economicconditions that immediately confronted his admin-istration.

Yet even in light of these circumstances,Roosevelt remained concerned about the debt beingincurred by his recovery programs. Building onthis, historians who claim that Roosevelt supportedthe reserve fund argue that he did so out of recogni-tion that payroll tax surpluses could be invested inthe public debt, thereby reducing net demands onthe Treasury. That Roosevelt was aware this couldbe done is clear—Morgenthau gave testimony tothis effect before the Ways and Means Committee—although it remains uncertain that this awarenesscontributed to his insistence on raising the taxrates proposed in the initial CES report.

Thus various explanations have been ad-vanced to explain Roosevelt’s allegiance to a fully


36

self-financed old-age insurance program. Asdiscussed earlier, one of these was his belief thatcontributory financing would politically secure theviability of old-age insurance by convincing indi-viduals that they had an “earned right” to benefitsand thus drawing their allegiance to the programby providing them with a sense of participation.Related to this, it is well known that Roosevelt hada canny sense of the cyclical nature of politics. Hewas aware that, although circumstances provided apropitious political climate for innovative liberalpolicy in the 1930s, these circumstances wouldeventually change, and with them, so to wouldpolitical proclivities. When this occurred, Rooseveltwanted to ensure that the old-age insuranceprogram would not be identified as “emergencylegislation” and consequently deemed no longernecessary. Avoiding this, Roosevelt was convinced,required the establishment of a program that wasboth self-supporting and strongly backed by thevoting public. In his view, contributory financingcould accomplish both these goals.

However, opposition to full contributoryfinancing came from two camps. One argued thatpayroll taxes were overly regressive and the otherthat economic growth would be impeded by thelarge reserve fund such a system would create.Those concerned with the regressivity of contribu-tory financing claimed that employers could passthe full burden of payroll taxes along to workers,either in the form of higher prices or lower wages.Referencing this position, several key political andsocial figures called for a system that would paygreater heed to tax equity. Notable among thesewere cabinet official Harry Hopkins, Senator andfuture Supreme Court Justice; social insurancetheorists Isaac Rubinow and Abraham Epstein; andtwo highly respected Congressmen, HenryEllenbogen (D-PA) and Rexford G. Tugwell (D-NY).According to Epstein, in testimony to the HouseWays and Means Committee, full contributoryfinancing would yield, “a system of compulsorypayments by the poor to the impoverished thatrelieves the well-to-do from their share of the socialburden” (Lubove, 1968).

To remedy this, those alarmed by the per-ceived regressiveness of the program advocated oneof two solutions. The first called for exemptingpersons in low-income groups from the payroll tax.The advantages of this were summarized by E. J.

McCormack, who served as a special assistant tothe Social Security Administration’s predecessor,the Social Security Board, in 1936. According toMcCormack, “Fifty cents a month (the averageamount low-income workers would save if ex-empted) is equivalent to four plus quarts of milk,and that much milk during the month to an infantmight prove a better investment in human valuesthan the same amount put away over a period ofyears to provide for the old-age of that infant’sfather” (Cates, 1983).

The second solution proposed by opponents ofcontributory financing called for a tax increase onbusiness and the wealthy in order to fund old-ageinsurance with general revenues. According toEllenbogen:

There is one defect in this old-age insurancesystem as set up in the act of Congress, avital, a fundamental defect. The FederalGovernment does not contribute to it. TheFederal government, as in many Europeancountries, should contribute one-third of thetotal fund. Where will it get the money? I donot want to use this speech as a springboardfor a dissertation on the maldistribution ofwealth and income in this country, but I willventure to state that in a country where87 percent of the wealth is owned by4 percent of its population, inheritance andincome taxes could well be increased for thispurpose3 (Congressional Record, 1935).

Ultimately, however, although this oppositioncreated a sense of urgency that expedited action onthe CES proposal, it failed to effect the structureand provisions of the Social Security Act. In addi-tion to Roosevelt’s concerns about the long-termpolitical viability of the program, scholars offer abroader explanation for this failure. At the time,federal taxing powers were extremely limited. Most

3 Proposals that reflected the Ellenbogen’s sentimentincluded the Lundeen Bill and Sen. Huey Long.’sShare-Our-Wealth campaign. Both plans calledimposing considerable taxes on wealth and businesses,and redistributing this money to the aged and unem-ployed. In March 1935, the Lundeen Bill passedthrough the House Labor Committee, and Sen. Longhad gained enough support nationally to concern theWhite House that his entrance into the 1936 presiden-tial election as a third party candidate could jeopar-dize Roosevelt’s chances for reelection.

37

Executive Summary

federal revenue came from tariffs or sales tax onnon essential items such as furs and cigarettes.Although an income tax had been instituted earlierin the century, by the time of the Depression, itremained highly underdeveloped. In fact, in theearly 1930s, over 95 percent of the U.S. populationwas exempt from all taxes. Thus to subsidize aprogram as massive as old-age insurance withgeneral revenues would have required a completeoverhaul of the tax system and, notably, an in-crease in direct taxation.

According to some scholars, the logistics ofaccomplishing this, coupled with the fact thatraising direct taxes in the midst of a depressionwould be unpopular at best, precluded significantreliance on general revenues. Full contributoryfinancing was also opposed by those who fearedthat the system’s large reserve fund dangerouslyprivileged savings over consumption. From theirperspective, excessive savings were already placinga massive drag on the American economy, and thekey to ending the Depression was growth spurredby consumption.

Expert testimony critical of the reserve wasgiven before the House Ways and Means Commit-tee and in the Senate, but no proposal for reducingthe reserve was offered. It was not until two yearsafter the act was passed that the issue of the old-age insurance reserve gained prominence, ulti-mately leading to major revisions of the SocialSecurity system passed in 1939.

■ The Reserve Fund and the1939 Amendments

Following the Social Security Act’s passage in 1935,it continued to face considerable challenges. Thelogistics of starting up the new system wereimmense; many businesses resisted compliance; theconstitutionality of the act was unresolved; andmany groups, most notably the Townsend Move-ment, continued to push for more generous ben-efits. Compounding these problems, a severerecession hit the country in 1937, and observersseeking to explain this downturn quickly pin-pointed the reserve fund, which had alreadywithdrawn approximately $2 billion in payrolltaxes from the economy. Calls mounted to addressthis accumulation of reserves, and Sen. ArthurVandenburg (R-MI), who said of the reserve fund “it

is scarcely imaginable that rational men shouldpropose such an unmanageable accumulation offunds in one place in a democracy,” called forreducing the reserve by liberalizing benefits andinstituting a payroll tax freeze (Tynes, 1996). It washoped that this increase in benefit payments anddecrease in taxes would stimulate the economythrough encouraging consumption. Responding toVandenburg’s proposal and accompanying pres-sures, Roosevelt formed the first ever AdvisoryCouncil on Social Security, charging them with thetask of suggesting reforms that would defuseattacks on the program.

Ironically, the council’s final report, issued onDecember 10, 1938, called for changes that practi-cally eliminated the effects of the MorgenthauAmendment. The report called for accelerating theinitial benefit payment schedule, increasing benefitamounts as well as the covered population, andfreezing tax rate increases. In the amendments of1939, passed with only10 dissenting votes, each of these recommendationswas enacted: the date of beginning payments wasmoved up to 1940; retirement payments in the firstyears of the program were increased; coverage wasextended to the survivors of both active and retiredemployees who had died and to the dependents ofretired workers; and the tax rate was frozen at1 percent for a period of 11 years.

These changes pleased a wide spectrum of thepopulation, appealing to business through the taxfreeze and workers through benefit increases andexpansions, but they also seriously altered two ofthe program’s original tenets. First, the contractualinsurance principle—i.e., the close linking ofbenefits and contributions—was dramaticallyweakened. With benefit rates for early recipientsnow raised above 5 percent and taxes frozen at1 percent until 1949, the program became lessactuarially sound than it had been in the originalCES plan rejected by Roosevelt. Second, shiftingfrom a reserve funded system to a pay-as-you-gosystem opened up the possibility that futuregenerations would be forced to meet benefit claimseither by raising taxes or dipping into generalrevenues. This prospect, particularly in light of thewillingness of Congress in 1939 to raise benefitsbut not taxes, alarmed many Social Securitysupporters. In his diary, Morgenthau feared that


38

the revisions “may be a device that willeventually kill Social Security,” and Witte foundhimself with “a sinking feeling about the future ofold-age insurance” (Leff, 1983).

■ Retirement Age and RelatedIssues

Although the topic did not receive significantattention during the formation of Social Security,alternatives to the benefit eligibility age of 65 weresuggested. There was considerable support, espe-cially from labor leaders, for setting the age at 60.They argued that the nature of industrializationhad accelerated the rate at which a human becameincapable of productive work. Consequently, theysuggested that the act disregard the precedent ofage 65 established in Europe and provide fullbenefits at age 60. Reinforcing their arguments,proponents of age 60 noted that life expectancy atbirth in 1929 was only 57.7 years for males. Accord-ing to Ellenbogen, “The 65 year limit…must go. Itis entirely too high. After all, this is supposed to bea pension for old-age, not a graveyard pension”(Congressional Record, 1935).

Several officials from the CES and theTreasury Department, concerned with long-rangefinancing, advocated age 70. They were stronglyopposed to making benefits available at age 60,arguing that this would greatly imperil thesystem’s capacity to be self-supporting. It is gener-ally agreed that the selection of age 65 was acompromise between those advocating age 60 andthose advocating age 70. Of greater interest wasthe question of whether or not completely leavingthe labor force should be a condition of receivingbenefits.

Two historical interpretations have emergedconcerning the retirement provisions. The firstsuggests that moving older workers out of the laborforce in order to make room for younger ones wasonly a minor motivation of the act. Supporters ofthis position note that the original act contained noprovision mandating retirement at age 65. In fact,in the version of the bill that passed through theHouse, individuals became eligible for full benefitsat age 65 regardless of whether or not they contin-ued to work. Eventually, through an amendmentadded in the Senate and approved in conference,this was altered so that individuals who continued

working at age 65 would not become eligible forbenefits, although they could defer benefits andcontinue working.

The historical record pertaining to thisrevision tends to indicate that the basis of theSenate amendment was actuarial and not due to adesire to improve employment conditions. In the1939 amendments to the act, the “retirement test”was modified to allow persons to continue workingand receiving old-age insurance benefits as long astheir earnings from “covered” employment did notexceed $15 per month.

Conversely, some historians suggest thatproviding encouragement for elderly workers to exitthe labor force was a significant motivation for theSocial Security Act. According to these scholars, theprime component of this motivation was a desire toundercut radicalism by providing the unemployedand disenfranchised with gainful employment.

It is clear that one of the goals of socialinsurance is the reduction of social turmoil. In theopinion of Wilbur Cohen, this was the foundingbasis of social insurance, whose “roots came out ofthe work and consideration of the people in thefield of labor legislation. Social insurance to themwas a form of remedial legislation to deal with theproblems of labor unrest in an industrial society”(Domhoff, 1990). Old-age insurance would address“labor unrest” by retiring older workers andconverting them into consumers through theprovision of a pension. In turn, jobs would immedi-ately be created as spending spurred economicgrowth. Illustrating this principle, Rep. ReubenWood (D-MO), drawing an analogy to pendinglegislation to begin a federal pension program forrailroad workers, stated, “If it should go into effect,it is estimated that in the first year it will take outof service approximately 250,000 railroad men,placing them on a pension or annuity. That wouldnaturally make openings for 250,000 younger men”(Tynes, 1996).

Roosevelt himself seemed sympathetic to thisrationale, emphasizing that the key to recovery wasemployment and work, not relief. It was his hopethat the assistance- based provisions of the billwould eventually recede in importance, and hebelieved that reducing unemployment was essen-tial for this to occur. Thus, in consultation with theCES, he considered including a mandatory retire-ment provision in the original bill. However this

39

Executive Summary

was never done, and a serious attempt to do sowould not again surface during the act’s inception.

■ The Maturing of Social SecuritySocial Security is one of our nation’s most impor-tant social programs. However, the growth of theprogram, and perhaps even its existence, remaineduncertain until 1950. Throughout the 1940s, old-age insurance was outpaced by public welfareprovisions, both in terms of beneficiaries andbenefit levels. In this climate, interest in theprogram waned, and its very necessity was ques-tioned by many. On several occasions, Congresscanceled scheduled increases in tax rates, reason-ing that there was little demonstrated need forprogram expansion. OAI was further hamperedwith the outbreak of World War II, which pusheddomestic issues into the political background. Afterthe war, when a depression did not materialize asmany feared, suggestions began to mount thatmeans-tested programs could sufficiently addressthe issue of old-age security.

Despite these unfavorable circumstances,Social Security was expanded in 1950. Benefitsincreased by 77 percent. The tax level was raised to3 percent, with both employer and employeecontributing 1.5 percent, and the taxable wage basewas raised from $3,000 to $3,600. In addition, thepayroll tax was extended to include self-employedworkers, who were taxed at a rate of 2.25 percent oftaxable payroll. Several factors contributed to thisexpansion, with two being particularly significant.First, the actions of an Advisory Council convenedin 1948 persuaded legislators that elderly depen-dence on the dreaded public “dole” could only belessened through the expansion of OAI. Here, thesame arguments used to ensure the act’s originalpassage—developing a system that linked benefitsand contributions—were again invoked. Second,beginning in the late 1940s, several labor unionspressured for and received private pension plans.Benefit levels in these employer-sponsored planswere set at a specific level and tied to SocialSecurity. For instance, the Big Three offered amonthly retirement pension of $100, with thecompany providing the difference between thisamount and the OAI benefit level. As a result ofthis link, employers could decrease their contribu-tions in proportion to OAI benefit increases.

Consequently, many became interested in expand-ing Social Security, and several historians suggestthat this interest was crucial to the passage of the1950 amendments.

It is generally agreed that the 1950 amend-ments “saved” Social Security, placing it on a pathof steady incremental expansion. Shortly after theamendments were passed, the number of personsreceiving OAI surpassed the number of old-ageassistance beneficiaries, and in August 1951, OAIpaid out more in benefits than old-age assistancefor the first time in U.S. history. OAI continued toexpand throughout the 1950s, benefiting fromsteadily rising wages and favorable worker-to-beneficiary ratios. Combined, these factors allowedlegislators to raise benefits without raising payrolltaxes. Coverage too was expanded, with agricul-tural workers incorporated into the system in 1953.

That much of the program’s growth occurredwith the blessing of a Republican President(Eisenhower) reflected the bipartisan support forOAI that had emerged by the mid-1950s. Theparties were clearly not in complete unison—asindicated by the experience of Disability Insurance(DI), narrowly enacted in 1956 after contentiousdebate and staunch Republican opposition—butboth were committed to the general principles andexistence of OAI, something that had not been thecase less than 10 years earlier. Additionally, publicopinion polls recorded that OAI was one of the mostpopular government programs. Thus both in termsof political and public support, Roosevelt’s desirefor a politically “untouchable” old-age insuranceprogram was reaching fruition.

Social Security continued its growth in the1960s, with OASDI benefit levels being increasedin both 1968 and 1969. In 1965, the Social SecurityAct was expanded to include health insurance forOASDI beneficiaries. Known as Medicare, thisamendment was the product of extensive compro-mises between advocates and opponents of compre-hensive national health insurance, and it builtcarefully on traditions well established by the OldAge Survivors Insurance program. For instance,eligibility for Medicare was contingent uponeligibility for OASDI; both programs were predomi-nantly financed via payroll taxes; and both pro-grams were portrayed as “earned” rights, attainedthrough contributions made during one’s


40

working years.Beginning in the 1970s, policy and rhetoric

concerning Social Security shifted away fromgradual expansion and toward preservation andsustainability. This shift was, and is, a product ofchanging demographic and economic circum-stances. In a pay-as-you-go retirement program,the ability to raise benefits more rapidly thanpayroll taxes is contingent upon two factors:favorable worker-to-beneficiary ratios and steadilyrising wages. Throughout the 1950s and 1960s,both of these conditions existed. However, in theearly 1970s, this began to change. Economicsstruck first, as productivity declined, driving pricesupward and wages downward. This had dualnegative effects on Social Security because benefitincreases had been indexed to the rate of inflationin 1972. As a result, rising prices automaticallyresulted in rising benefits. However, wages werenot keeping pace. As a result, OASDI was saddledwith increased liabilities (higher benefit payments)and decreased revenue opportunities (a decreasedtax base).

By the early 1980s, this bifurcation betweenwage increases and benefit increases was growingrapidly, with many suggesting that the trend wasunsustainable. For instance, benefits automaticallyincreased by 14.3 percent in 1980, but wages roseonly 9 percent. Alarmed by these conditions Presi-dent Reagan charged a bipartisan commission withthe task of resolving the crisis. This ultimately ledto several revisions in the Social Security program,including the taxation of some benefits, the inclu-sion of new federal employees in the OASDIsystem, a six-month delay in automatic cost-of-living adjustments, and small payroll tax increases.Additionally, the normal retirement age was raised,with an increase to age 67 beginning in 2003.

These revisions largely addressed the financ-ing dilemmas confronting OASI in the early 1980s.In 1983, the OASI Trust Fund balance stood at$19.7 billion. Today it is over $440 billion. Yet,although the program’s short term finances areensured, pending demographic changes threaten itslong-range solvency. Additionally, unfavorabledemographics are undermining current andprojected rates of return. As the number of benefi-ciaries grows more rapidly than that of covered

workers, the OASI tax base declines and its obliga-tions increase. For instance, in 1960, a singleretiree could expect to receive benefits betweenseven and nine times higher than his or her totalpayroll contributions. By 1980, this favorable rateof return had declined to two to three times therate of contribution, and projections suggest thatworkers retiring in 2025 will receive benefitpayments equal to only 75 percent to 90 percent oftheir total lifetime contributions. Currently, thereare 3.2 workers per beneficiary. In 2025, it isprojected that there will be 1.9 workers per benefi-ciary.

Traditionally, Social Security finances areviewed by analysts over a 75-year period.4 Due tounfavorable demographic trends and continuedrates of slow productivity, OASDI is underfundedover such a period by 2.17 percent of coveredpayroll. Combined with trends that clearly indicatediminishing rates of return for OASDI participants,this level of underfunding has prompted many tosuggest that Social Security is in need of funda-mental reform. Two suggestions that would signifi-cantly alter the program have emerged—privatization and “affluence” testing.

The passage of Social Security meant that, forthe first time, the federal government was in thebusiness of organizing social insurance. Prior tothis, individual effort and family support were thesole guarantors of financial security in old age.However, in shaping the Social Security Act,policymakers paid careful attention to the ethos ofindividual effort—i.e., they tied effort to reward byrelying on contributory financing. In this sense thephilosophical underpinnings of Social Securitycontain a healthy respect for private initiative andindividual effort. But the rate of return to peoplewith careers marked by higher wages is compro-mised in favor of supporting those with lowerlifetime earnings. These low rates of return, inconjunction with unfunded liabilities in the future,bring us right back to the struggle to define the roleof social insurance in our market-based economy.

■ Bibliography

4 Seventy-five was selected because it approximatesone lifetime.

41

Executive Summary

Achenbaum, Andrew W. “Mandating: What Can WeLearn From History?” In Government Mandatingof Employee Benefits. Washington, DC: EmployeeBenefits Research Institute, 1987.

Achenbaum, W. Andrew. Social Security: Visions andRevisions. Cambridge, UK: Cambridge Univer-sity Press, 1986.

Cates, Jerry. Insuring Inequality: AdministrativeLeadership in Social Security, 1935–54. AnnArbor, MI: University of Michigan Press, 1983.

Cohen, Wilbur J., and Robert Ball, eds. The Report ofthe Committee on Economic Security of 1935,50th Anniversary Edition. Washington, DC:National Conference on Social Welfare, 1985.

Congressional Record. Congressional Record. Vol. 79,74th Congress, First Session. Washington, DC:Government Printing Office, 1935.

Derthick, Martha. Policymaking for Social Security.Washington, DC: Brookings Institution, 1979.

Domhoff, William G. The Power Elite and the State:How Policy Is Made in America. New York:Aldine De Gruyter, 1990.

Friedland, Robert B. When Support and ConfidenceAre at Odds: The Public’s Understanding of theSocial Security Program, Washington, DC:National Academy of Social Insurance, 1994.

Graebner, William. A History of Retirement: TheMeaning and Function of an American Institu-tion. New Haven, CT: Yale University Press,1980.

Haber, William, and Wilbur J. Cohen. Social Secu-rity: Programs, Problems, and Policies.Homewood, IL: Richard D. Irwin, Inc., 1960.

Jacobs, Lawrence R. The Health of Nations: PublicOpinion and the Making of American andBritish Health Policy. Ithaca, NY: CornellUniversity Press, 1993.

Jenkins, J. Craig, and Barbara Brents. “SocialProtest, Hegemonic Competition, and SocialReform: A Political Struggle Interpretation of theOrigins of the American Welfare State.” Ameri-can Sociological Review 54 (1989): 891–909.

Kingson, Eric R., and Edward Berkowitz. SocialSecurity and Medicare: A Policy Primer.Westport, CT: Auburn House, 1993.

Kingson, Eric R., and James H. Schulz, eds. SocialSecurity in the 21st Century. New York: OxfordUniversity Press, 1996.

Latimer, Murray W. Industrial Pensions in theUnited States and Canada. New York: IndustrialRelations Counselors, 1932.

Leff, Mark. “Taxing the ‘Forgotten Man:’ The Politicsof Social Security Finance in the New Deal.”

Journal of American History. Vol. 70 (1983):359–381.

Leuchtenberg, William E. Franklin D. Roosevelt andthe New Deal, 1932–1940. New York, NY:Harper and Row, 1963.

Lubove, Roy. The Struggle for Social Security, 1900–1935. Cambridge, MA: Harvard UniversityPress, 1968.

Orloff, Ann Shola. The Politics of Pensions: AComparative Analysis of Britain, Canada, andthe United States, 1880–1940. Madison, WI: TheUniversity of Wisconsin Press, 1993.

Quandagno, Jill S. “Welfare Capitalism and theSocial Security Act of 1935.” American Sociologi-cal Review 49 (1984): 632–647.

Reno, Virginia, and Robert B. Friedland. “StrongSupport But Low Confidence: What Explainsthe Contradiction?” Eric Kingson and JamesSchultz, eds., Social Security in the 21st Cen-tury, New York, NY: Oxford University Press,1997.

Roosevelt, Franklin D. “Addresses to the AdvisoryCouncil on Social Security” (November, 14,1934). In Samuel Rosenman, The Public Papersand Addresses of Franklin D. Roosevelt. Vol. 3.New York, NY: Random House, 1938.

Schieber, Sylvester J. A Proposal to EstablishPersonal Security Accounts as an Element ofSocial Security Reform as Considered by theSocial Security Advisory Council. TestimonyBefore the Senate Finance Committee,March 25, 1996.

Schiltz, Michael E. Public Attitudes Toward SocialSecurity: 1935-1965. Washington, DC: SocialSecurity Administration, 1970.

Skocpol, Theda, and Edwin Amenta. Did CapitalistsShape Social Security?” American SociologicalReview 50 (1985): 572–575.

Stevens, Beth. “Blurring the Boundaries: HowFederal Social Policy Has Influenced WelfareBenefits in Private Sector.” In Ann Orloff,Margaret Weir, and Theda Skocpol, eds., ThePolitics of Social Policy in the United States.Princeton, NJ: Princeton University Press,1988.

Tynes, Sheryl R. Turning Points in Social Security:From “Cruel Hoax” to “Sacred Entitlement.”Stanford, CA: Stanford University Press, 1996.

Witte, Edwin E. The Development of the SocialSecurity Act. Madison, WI: The University ofWisconsin Press, 1963.

43

Chapter 3

EBRI Social Security ReformAnalysis Project Progress Report:Phases 1 and 2

3by Martin R. Holmer

■ PrefaceThis report was produced as part of the EmployeeBenefit Research Institute (EBRI) Social Securityreform project. (A previous version of this reportwas presented at the EBRI Policy Forum on“Assessing Social Security Reform Alternatives’’ inDecember 1996.) While the model developmenteffort was not complete, it had progressed farenough to provide some unique insights into thenature of the Social Security problem and thebenefits, cost, and risks of alternative approachesto solving the problem. Subsequent work under theEBRI Social Security reform project will produce amore complete narrative of the results outlinedhere.

■ Understanding via QuantitativeAnalysis

This section describes the role of quantitativeanalysis in understanding the nature of the SocialSecurity problem and in assessing the benefits,costs, and risks of alternative proposals for reform-ing current Social Security policy.

First, a short account of the recent broadeningof the reform debate shows why a new kind ofpolicy simulation model is needed to conduct a fairand complete analysis of the structural reformsthat have been proposed in the past few years.Models that were adequate for analyzing thenonstructural reforms that dominated the earlierdebate are not designed to address many of the keyissues raised by the new structural reform propos-als.

Second, a short description of the EBRI-SSASIM2 stochastic policy simulation model isprovided. This model is being developed to enable afair and complete comparison of the benefits, costs,and risks of a wide range of Social Security reformproposals.

■ Rationale for a New Kind ofModel

A major premise of the EBRI Social Security reformproject is that better understanding of the problemscurrent-law Social Security policy faces requiresquantitative analysis that estimates all the ben-efits, costs, and risks of the program. Likewise,quantitative analysis of all benefits, costs, andrisks is required for a fair and complete comparisonof the broad range of reform proposals that arebeing offered as solutions to these problems.

The emphasis on quantitative analysis isrooted in the belief that any careful assessment of aproposed change requires estimates of the effects ofthat change. Because the effects of changes inSocial Security policy will take many decades tobecome apparent, computer simulation is the onlyviable method of estimating the effects of change.Stochastic simulation is the appropriate analysismethod because the effects of a policy change oftendepend on the size and composition of the popula-tion and on the state of the economy, neither ofwhich is known with certainty in the future. Use ofMonte Carlo simulation methods permits quantita-tive estimates of how the effects of policy reformwill vary depending on the state of the economyand the size and composition of the population.

The scope of current reform proposals is muchbroader than in the past. For example, some reformproposals call for equity investment of the trustfunds, while others call for the creation of self-managed personal retirement accounts. The rangeof analysis issues raised by such major structuralreforms is much broader than those raised by thenonstructural reforms that have been analyzedover the past several decades.

This broadening of the scope of analysis hascreated demand for new kinds of quantitativeanalysis tools. Policy simulation modeling has a


44

long tradition in the Social Security policy analysiscommunity, but the kinds of models currently inuse are designed to analyze the nonstructuralreforms that have dominated debate in the past.These older models are not well suited to analyzemany of the new issues raised by the structuralreform proposals.

Many of the shortcomings of current modelsstem from the fact that they are not designed toanalyze defined contribution-style personal ac-counts or risky investments. But having quantita-tive estimates of both average results and thevariability of those results is essential to anyevaluation of a policy’s benefits, costs, and risks.

In addition, current models fail to provide anadequate method of characterizing the economicfeedback effects of Social Security policy. Reformproposals that call for major changes in retirementages and/or introduction of personal retirementaccounts are likely to affect aggregate work effortand national saving in ways that alter the patternof economic growth, which in turn can affectprogram tax revenues and benefit costs.

Considering this broadening in the scope ofSocial Security policy debate, it is clear that thereis a need for a new kind of policy simulation modelthat can provide quantitative estimates for the fullrange of benefit, cost, and risk issues that are nowbeing addressed in discussions of alternative policyreform proposals. Current modeling technology,which is adequate to support yesterday’s policydiscussion, needs to progress to support today’sdiscussion, which focuses on the effects of major,structural reforms in Social Security policy.

EBRI-SSASIM2 Model Overview

The model is a dynamic and stochastic policysimulation of the current Social Security system,reforms in that defined benefit system, and propos-als for introducing defined contribution featuresinto the system. The model extends in numerousways the capabilities of the original SSASIMmodel, which was developed for the Social SecurityAdministration’s Advisory Council. The basiccapabilities of the original model and the newmodel’s major enhancements are described in turn.

The SSASIM model was originally developedto explore the use of Monte Carlo simulationmethods to characterize the demographic andeconomic uncertainty facing the Social Security

program. Exploration of this stochastic simulationmethod was a major recommendation of the priorTechnical Panel on Assumptions and Methods, aswell as the panel appointed by the current AdvisoryCouncil. Initial demographic modeling results werereported in the panel’s final report. The AdvisoryCouncil sponsored additional model development,which added economic, asset return, tax, andsimplified benefit modules, to enable the model toanalyze the risks as well as the benefits and costsof equity investment options in the trust funds.

After this initial stage of model development,the model’s logical structure, including its inputassumptions and output results, was similar to thatof the Social Security Administration’s ActuarialModel. The major differences were that program-related risks were explicitly represented usingMonte Carlo methods and that several economicfeedback effects were designed into the model.Monte Carlo methods are used to characterizeuncertainty about the future course of 13 keydemographic and economic input variables used inthe Social Security Actuarial Model, as well asuncertainty about future asset returns. Operatingin a nonstochastic mode, the model closely repli-cates each of the three scenarios presented in theTrustees’ Report.

The range of reform proposals currently beingdiscussed requires an enhanced model to assessfairly the benefits, costs, and risks of differentproposals. The EBRI-SSASIM2 model buildsdirectly on the earlier model and adds majorenhancements in a number of important areas:• Implementation of structural benefit modules

that enable more detailed analysis of a widerange of benefit reforms in the current definedbenefit Social Security structure.

• Implementation of economic feedback linkages,including a saving, investment, and productivitygrowth linkage and an asset-allocation, relativeasset returns linkage.

• Development of cohort lifetime experienceanalysis capabilities that produce policy perfor-mance indicators (such as money’s worth return)from the same model and assumptions thatproduce aggregate program financial results anduse realistic age-earnings’ profiles.

• Development of cohort policy performanceindicators that measure not only the averageprogram experience of a cohort and different

45

Chapter 3

cohort subgroups but also measure the risksthat alternative policies impose on differentcohort groups.

• Development of personal retirement accountmodules that represent the basic features of arange of current defined contribution reformproposals and characterize the range of accountbalance, asset-allocation, and annuitizationbehavior.

• Use of stochastic asset returns to model realisti-cally the pattern of accumulation of personalretirement account balances.

■ Examining the Nature of theProblem

This section begins the examination of the futurecost of current-law Old-Age and Survivors Insur-ance (OASI) benefit policy by looking at resultsfrom the best-guess, single-scenario projection usedin the Trustees’ Report. It continues by establishingthe demographic roots of current-law benefitpolicy’s long-run actuarial deficit. Next, it examinesthe uncertainty of future costs, using assumptionsabout future variability in the demographic andeconomic environment of the program that areconsistent with those assumed in the Trustees’

Year

Perc

enta

ge o

f Tax

able

Pay

roll

-2

0

2

4

6

8

10

12

14

16

2000 2010 2020 2030 2040 2050 2060 2070

OASI Cost Rate

Income Rate

Actuarial Deficit

Note: The solid line represents the OASI cost rate, the broken line the income rate, and the dot-dash line represents the actuarialdeficit (the negative of the Trustees’ Report’s balance). Estimates are from model run 202, which generates one scenario.

Chart 3.1Cost Problem for Current-Law Benefit Policy Under Trustees’ Report

Intermediate-Cost Assumptions

Report. Then it estimates the cost effects of chang-ing only the rate of mortality decline from theassumption used in the Trustees’ Report to anassumption that is closer to that of the CensusBureau. The section concludes with a discussion ofthe size and timing of the OASI cost problem.

Future Cost of Current-Law Benefit Policy

The cost of OASI benefit policy will rise sharply asbaby-boom cohort members begin to retire. Thecurrent level of payroll tax rates will not be suffi-cient to finance these higher costs. See chart 3.1 fordetails.

The cost problem remains even after mostbaby-boom cohort members have died in 2070.Chart 3.1 shows the gap between costs and incomein 2070.

The long-run average cost rate and long-runaverage actuarial deficit estimates understateextent of the cost problem because they includecurrent low costs and current program surpluses.Table 3.1 shows difference the between 2070 andlong-run average cost rates and deficits.

EBRI-SSASIM2 model, when running in non-stochastic, single-scenario mode with the sameintermediate-cost demographic and economicassumptions, produces estimates that are similar to


46

Table 3.1Summary Estimates of Cost Problem for Current-Law Benefit Policy under Trustees’

Report Intermediate-Cost Assumptions

Source of Long-Run Long-Run Cost Rate DeficitEstimates Cost Rate Deficit in 2070 in 2070

Model Run 202 13.26 1.84 15.91 4.44

1996 Trustees’ Report 13.33 1.85 16.39 4.91

Note: The long-run cost rate (deficit) refers to the Trustees’ Report’s 75-year summarized cost rate (negative actuarial balance). Allestimates are expressed in terms of percent of taxable payroll. [96TR, p.108 & p.113]

those in the 1996 Trustees’ Report. See table 3.1 fora comparison of these estimates. Not shown is theclose agreement between model and Trustees’Report estimates under both the low-cost and high-cost assumptions.

Demographic Roots of the Cost Problem

Life expectancy at birth has risen from less than64 years in 1940, when Social Security first paidbenefits, to over 76 years today. Even though theintermediate-cost assumptions used in the Trust-ees’ Report project a future rate of mortality declinethat is substantially below the average rate of the20th century, life expectancy is estimated tocontinue to increase, reaching age 81 by 2070

Year

Life

Exp

ecta

ncy

at B

irth

62

6466

68

70

72

7476

78

80

82

1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 2060 2070

Intermediate Cost Assumptions

Note: The solid line represents historical values before 2000 and Trustees’ Report estimates beginning in 2000. The broken linerepresents estimates using intermediate-cost assumptions from the single-scenario model run 202. In both cases, male andfemale life expectancy is averaged to obtain a combined life expectancy.

Chart 3.2Historical and Projected Life Expectancy at Birth under Trustees’ Report

Intermediate Cost Assumptions

(chart 3.2).Shorter life expectancy would eliminate the

Social Security cost problem as shown in table 3.2.This indicates that the future problem is rooted inthe relatively small adjustment in current-lawbenefit policy (i.e., no change in the early retire-ment age of 62 and a planned increase in thenormal retirement age from 65 to 67) in the face ofa substantial increase in life span.

Cost Uncertainty Using Trustees’ Views

Future values of demographic and economicvariables that influence OASI cost and incomecannot be accurately forecast over the next75 years. The Trustees’ Report recognizes this in an

Trustees Report

47

Chapter 3

Table 3.2Summary Estimates of Cost Problem for

Current-Law Benefit Policy Assuming

Shorter Life Expectancies

Assumed ConstantLife Expectancy Long-Run Deficit Deficit in 2070

76.28 1.18 2.7072.60 –0.09 1.1469.55 –1.12 –0.09

Note: All estimates assume no mortality decline from initialmortality rates that have been increased to simulate earlierhistorical life expectancies. The cost estimates are from modelrun 204, which is a nonstochastic, single-scenario run usingintermediate-cost assumptions except for mortality. Allestimates are expressed in terms of percent of taxable payroll.

Table 3.3Range of Cost Estimates for Current-Law Benefit Policy,

Using Three Sets of Trustees’ Report Assumptions

Long-Run Long-Run Cost Rate DeficitEstimate Source Cost Rate Deficit in 2070 in 2070

Model Run 201 11.13 –0.19 11.24 0.03TR Low-Cost 11.09 0.25 11.46 0.25

Model Run 202 13.26 1.84 15.91 4.44TR Intermediate-Cost 13.33 1.85 16.39 4.91

Model Run 203 16.06 4.52 23.24 11.39TR High-Cost 16.23 4.61 24.43 12.52

Note: The long-run cost rate (deficit) refers to the Trustees’ Report’s 75-year summarized cost rate (negative actuarial balance) forOASI. All estimates are expressed in terms of percent of taxable payroll. [96TR. DD.108-109 & D.113]

tic simulation that reflects the uncertainty aboutthe future values of demographic and economicvariables implicit in the Trustees’ Report. Ten ofthe thirteen major variables that vary across thethree Trustees’ Report scenarios are assumed tohave a constant ultimate value that is drawn froma normal distribution with a mean equal to theTrustees’ Report’s intermediate-cost assumptionand a standard deviation equal to one-fourth of thedifference between the high-cost and low-costassumptions. The long-run, ultimate value distri-butions of these 10 variables are assumed to beuncorrelated. The three other variables—theunemployment rate, inflation rate, and nominalinterest rate—are assumed to fluctuate aroundmeans equal to the Trustees’ Report’s intermediate-cost assumptions, with deviations from the long-run mean being generated by a second-order vectorautoregressive process that has been estimatedwith historical data from the late 1920s throughthe early 1990s. The errors terms of these threedeviation processes were found in the statistical

ad hoc way by presenting low-cost and high-costscenarios in addition to the intermediate-costscenario. Table 3.3 gives estimates on how the sizeof the cost problem varies across these threescenarios.

Table 3.3 also contains OASI cost estimatesproduced by the EBRI-SSASIM2 model using thethree sets of Trustees’ Report assumptions. Thiscomparison indicates that the two models producesimilar estimates when using the same assump-tions. There is wide agreement that this approachneeds to be improved so that the effect of futuredemographic and economic uncertainty on aggre-gate program cost and income can be determined.1

Trustees’ Report assumptions are used inEBRI-SSASIM2 model run 205 to specify a stochas-

1 See, for example, Richard S. Foster, “A StochasticEvaluation of the Short-Range Economic Assumptionsin the 1994 OASDI Trustees Report,’’ Actuarial StudyNo. 109 (Baltimore, MD: Social Security Administra-tion, Office of the Actuary, 1994); Ronald D. Lee andShripad Tuljapurkar, “Stochastic Population Forecastsfor the United States: Beyond High, Medium, andLow,’’ Journal of the American Statistical Association(December 1994); and the last two reports of thequadrennial Social Security Advisory Council’sTechnical Panel on Assumptions and Methods.


48

Prob

abili

ty

0

0.02

0.04

0.06

0.08

0.1

0.12

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

2070 Cost Rate

Long-Run Average Cost Rate

Percentage of Taxable Payroll

Chart 3.3Long-Run Average and 2070 Cost Rate Distribution under Stochastic Assumptions

Consistent with Trustees’ Report

Note: The long-run average Old-Age and Survivors Insurance (OASI) cost rate estimates have a mean of 13.94 and a standarddeviation of 1.02; the probability of an estimate exceeding 16 is 0.023. The 2070 OASI cost rate estimates have a mean of 17.04and a standard deviation of 2.26; the probability of an estimate exceeding 20 is 0.107. Estimates are from model run 205, whichgenerates 1,000 stochastic scenarios.


Prob

abili

ty

0

0.02

0.04

0.06

0.08

0.1

0.12

-4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Long-Run Average Actuarial Deficit

2070 Actuarial Deficit

Note: The long-run average Old-Age and Survivors Insurance (OASI) actuarial deficit estimates have a mean of 2.50 and a standarddeviation of 1.03; the probability of an estimate exceeding 4 is 0.059. The 2070 OASI actuarial deficit estimates have a mean of 5.52 anda standard deviation of 2.14; the probability of an estimate exceeding 10 is 0.037. Estimates are from model run 205, which generates1,000 stochastic scenarios.

Chart 3.4Long-Run Average and 2070 Actuarial Deficit Distribution under Stochastic Assumptions

Consistent with Trustees’ Report

49

Chapter 3

estimation to be contemporaneously correlated.Estimates of the degree of cost uncertainty

produced by recognizing uncertainty in the futurevalues of these 13 demographic and economicvariables are shown in chart 3.3 for the cost rateand in chart 3.4 for the actuarial deficit. Thevariability in these estimates shows the consider-able cost uncertainty facing the OASI program,which is roughly consistent with the assumptionthat the Trustees’ Report’s high-cost/low-cost rangerepresents a 95 percent confidence interval for thedemographic variables and the economic variablesother than the unemployment rate, inflation rate,and nominal interest rate.

Cost Uncertainty Using Census’ Mortality View

The Trustees’ Report assumes that the future rateof mortality decline will decrease well below theaverage rate over the 20th century. This is acontroversial assumption that has a major effect onestimates of the size of the Social Security costproblem. Some demographers are more comfortablewith an assumption that projects past mortalitydecline rates into the future, and others argue thata shift to healthier life styles will increase the rateof mortality decline. In addition, some believe thatfuture biomedical advances will lead to longer lifeexpectancy.

Census Bureau projections of the rate ofmortality decline are considerably higher than theTrustees’ Report intermediate-cost assumption. Thedifference between the Census high-range and low-range assumptions is much larger than the Trust-ees’ Report range. Table 3.4 compares differentassumptions about the future rate of mortalitydecline.

EBRI-SSASIM2 model run 215 is specified tobe exactly the same as run 205 (the stochastic runconsistent with Trustees’ Report assumptionsdiscussed above), except that the mean rate ofmortality decline has been increased to be consis-tent with the Census Bureau’s mid-range estimate.This implied 1.0 percent mortality decline rateassumption is the same as the assumption recentlyrecommended by the Advisory Council’s TechnicalPanel. The standard deviation of the rate of mortal-ity decline is twice that assumed in run 205, whichproduces somewhat less variability than indicatedby the Census Bureau’s high/low range. It was notpossible to reproduce the likelihood of rates ofmortality decline near the Census Bureau’s high-range estimate because the three Census estimatesare asymmetric. This interpretation assumes thatthe high/low range represents a 95 percent confi-dence interval.

Census Bureau assumptions imply a lifeexpectancy at birth in 2070 that is on averageabout three years longer than that implied by theTrustees’ Report mortality decline assumptions.The probability of life expectancy at birth in 2070exceeding age 85 grows from one-half percentunder the Trustees’ Report assumptions to37 percent under the Census Bureau assumptions.See chart 3.5 for a comparison of the distribution oflife expectancy at birth under the Trustees’ Reportassumptions and the Census Bureau assumptions.

Census Bureau mortality assumptions implya worsening of the Social Security cost problem.The mean of the long-run actuarial deficit risesfrom 2.50 percent of taxable payroll to 3.13 percent,with the probability of it exceeding 4 percent risingfrom 0.059 to 0.265. The mean of the annualactuarial deficit in 2070 rises from 5.52 percent oftaxable payroll to 7.05 percent, with the probabilityof it exceeding 8 percent rising from 0.121 to 0.331.Details of the cost implications of the Censusmortality assumptions are shown in chart 3.6 forthe cost rate and in chart 3.7 for the actuarialdeficit.

Note: Census Bureau estimates were published in February1996. Research by sources marked with an asterisk (*) havebeen supported by National Institute on Aging. Estimatesmarked with a diamond (◆) are produced by the EBRI-SSASIM2model; all others are drawn from a National Institute on Agingchart.

Estimate Source Population 85+

Actual 1994 3.3Census Bureau (low) 9.6Olshansky* 11.4Trustees’ Report (low)◆ 11.8Trustees’ Report (intermediate) 14.6Trustees’ Report (high)◆ 17.8Census Bureau (mid) 18.2Lee* 21.4Census Bureau (high) 31.1Vaupel* 39.0Manton* 48.7

Table 3.4Alternative Estimates of Population

Ages 85 and Older in 2050

(All estimates are expressed in millions of people)


50

Prob

abili

ty

0

0.02

0.04

0.06

0.08

0.1

0.12

65 67 69 71 73 75 77 79 81 83 85 87 89 91 93

Life Expectancy in Years

Trustees’ Report

Census Bureau

Chart 3.5Distribution of Life Expectancy at Birth in 2070 under Stochastic Assumptions

Consistent with Trustees’ Report and Approximating Census Bureau’s

Note: The estimates in the Trustees’ Report distribution have a mean of 80.49 and a standard deviation of 1.88; the probabilityof an estimate exceeding 85 is 0.005. The estimates in the Census Bureau distribution have a mean of 83.41 and a standarddeviation of 4.30; the probability of an estimate exceeding 85 is 0.370. Estimates are from model run 205 (Trustees’ Report)and model run 215 (Census), both of which generate 1,000 stochastic scenarios.

Note: The long-run average Old-Age and Survivors Insurance (OASI) cost rate estimates have a mean of 14.61 and a standard deviationof 1.43; the probability of an estimate exceeding 16 is 0.169. The 2070 OASI cost rate estimates have a mean of 18.66 and a standarddeviation of 3.24; the probability of an estimate exceeding 20 is 0.288. Estimates are from model run 215, which generates1,000 stochastic scenarios.


Prob

abili

ty

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

2070 Cost Rate

Long-Run Average Cost Rate

Chart 3.6Long-Run Average and 2070 Cost Rate Distribution under Stochastic Assumptions

Consistent with Trustees’ Report Except for Use of Mortality Decline Assumptions

Approximating Census Bureau’s

51

Chapter 3


Prob

abili

ty

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

-4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Chart 3.7Long-Run Average and 2070 Actuarial Deficit Distribution under Stochastic Assumptions

Consistent with Trustees’ Report Except for Use of Mortality Decline Assumptions

Approximating Census Bureau’s

2070 Actuarial Deficit

Note: The long-run average Old-Age and Survivors Insurance (OASI) actuarial deficit estimates have a mean of 3.13 and a standarddeviation of 1.41; the probability of an estimate exceeding 4 is 0.265. The 2070 OASI actuarial deficit estimates have a mean of 7.05 anda standard deviation of 3.06; the probability of an estimate exceeding 10 is 0.154. Estimates are from model run 215, which generates1,000 stochastic scenarios.

Size and Timing of the Cost Problem

Cost and deficit estimates expressed as a percent oftaxable payroll may seem small until they areconverted to dollar amounts, which is the standardway of expressing most government program costsand the federal government’s overall deficit.

Taxable payroll was about $2,920 billion in1995. This means that 1 percent of taxable payrollis about $29 billion in 1995. It also means that adeficit of 7 percent, which is the mean 2070 OASIdeficit using the Census mortality assumptionsalong with other Trustees’ Report assumptions, isequivalent to $203 billion. This dollar deficitcompares with a federal government deficit of about$164 billion in 1995.

The size of the problem is large from otherperspectives as well. The Social Security deficitestimates can be translated directly into payroll taxincreases, which directly increase overall marginaltax rates for most workers. The size of the SocialSecurity cost problem, along with the Medicare costproblem, is large enough to produce a reversal ineconomic growth sometime in the first half of thenext century if neither problem is solved.2

The timing of the Social Security cost problemis conventionally measured by two milestones: theyear the trust fund ratio reaches its peak value(which is closely related to the year during whichthe current annual Social Security surplus turnsinto a deficit) and the year during which the trustfund balance reaches zero (which is the year of fundinsolvency).

The maximum OASI trust fund ratio will bereached most likely during the decade between2005 and 2015, with the chances of a later yearbeing only about 6 percent. After that time the fundwill begin to run a deficit and need to sell (not buy)Treasury bonds to finance costs. The medianestimate of 2010 compares with an intermediate-cost estimate of 2012 reported in the 1996 Trustees’Report. See chart 3.8 for the distribution of themaximum-fund-ratio year under assumptions

2 See the analysis of the economic feedback effects ofthe deficit financing of rising Social Security andMedicare costs in “The Long-Term Budget Outlook,’’ inThe Economic and Budget Outlook: Fiscal Years1997–2006 (Washington, DC: Congressional BudgetOffice, 1996).

Long-Run Average Actuarial Deficit


52

Chart 3.8Maximum-Fund-Ratio Year Distribution and Fund-Insolvency Year Distribution

under Stochastic Assumptions Consistent with Trustees’ Report Except

for Use of Mortality Decline Assumptions Approximating Census Bureau’s

Prob

abili

ty

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

1990 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060 2065 2070 2075

Year of OASI Trust Fund Insolvency

Note: Estimates of the maximum-fund-ratio year have a mean of about 2009 and a standard deviation of about 7; the probability of anestimated year being before 2005 is 0.170; the probability of an estimated year being after 2015 is 0.063. Estimates of the fund-insolvency year have a mean of about 2026 and a standard deviation of about 7; the probability of an estimated year being before 2020is 0.089; the probability of an estimated year being after 2035 is 0.047. Estimates are from model run 215, which generates 1,000stochastic scenarios.

consistent with Trustees’ Report assumptionsexcept for use of the Census Bureau mortalityassumptions.

OASI trust fund insolvency will occur mostlikely between 2020 and 2035, with the chances of alater year being only about 5 percent. The medianestimate of 2025 compares with an intermediate-cost estimate of 2031 reported in the 1996 Trustees’Report. See chart 3.8 for the distribution of thefund-insolvency year under assumptions consistentwith Trustees’ Report assumptions except for theuse of Census Bureau mortality assumptions.

■ Examining Alternative SolutionsThe previous section examined the nature of theSocial Security cost problem. This section examinesthe benefit, cost, and risk effects of alternativeapproaches to solving the cost problem.

In this preliminary examination the focus inon generic solutions rather than detailed reforms

that have actually been proposed in the publicdebate. The analysis of generic solutions will makeit easier to identify the relative strengths andweaknesses of widely differing approaches tosolving the cost problem. This comparative analysismay also suggest ways to specify more finelycrafted reform proposals that would ameliorate anyweaknesses identified in the generic solutions.

The range of possible nonstructural solutionsis bracketed by considering two generic approachesto reforming Social Security that leave Old AgeInsurance (OAI) as a pure defined benefit program.The first generic reform leaves OASI benefit policyunchanged and calls for a payroll tax rate thatgradually rises in future years to levels needed tofinance benefits on a pay-as-you-go basis. Thiscurrent-law benefits and increased taxes approachis referred to in this section as the CBIT reform.

The second generic solution leaves OASIpayroll tax rates at their current-law level andgradually reduces benefits in future years to levels

Year of Maximum OASI Trust Fund Ratio

Year

53

Chapter 3

that can be financed by taxes on a pay-as-you-gobasis. This reduced benefits and current-law taxesapproach is referred to in this section as the RBCTreform. These two solutions span the range ofnonstructural reforms that maintain the definedbenefit nature of Social Security and solve theprogram’s cost problem.

Structural reform of Social Security is repre-sented by the third generic solution examined inthis section. This approach to reform calls forestablishment of defined contribution (DC) stylepersonal retirement accounts with mandatorycontributions equal to 5 percent of taxable payrolland the gradual scaling back of the defined benefit(DB) portion of OAI. This reduction of the DBportion of OAI is specified so that DB benefitsdecline gradually over the 1999–2040 period asbirth cohorts retire having increasingly more yearsto accumulate account balances. During this period,the current payroll tax is maintained to finance thetransition cost from a pay-as-you-go program.During this transition period, current-law tax ratesare maintained in addition to the 5 percent accountcontribution rate. Beginning in 2040, the payrolltax is reduced so that the combined mandatoryaccount contribution rate and reduced payroll taxrate equals the current-law payroll tax rate.

This generic structural solution changes OAIfrom a pure defined benefit program to a mixedDC-DB program that relies heavily on the DCaccount to supply Social Security retirementbenefits. This approach is referred to in this sectionas the IARB reform because its main feature is tointroduce accounts and reduce DB benefits.

It is important to emphasize that the genericreform analysis presented below is preliminary inseveral ways. First, the reforms are quite stylizedbecause project time has been devoted primarily todeveloping the model’s analysis capabilities ratherthan to specifying more detailed proposals, and alsobecause it was felt that generic reforms are appro-priate for an initial comparison of widely differingapproaches to solving the Social Security costproblem. Second, it is possible that these prelimi-nary results may be revised as analysts gain moreexperience with the enhanced model. Third, someresults (such as the economic feedback effects ofreforms on the rate of national saving, investment,productivity growth, and real wage growth) will notbe available until after the next phase of modeldevelopment.

Current Benefits and Increased Taxes (CBIT)

CBIT generic reform maintains current-law OASIbenefit policy and gradually increases the OASIpayroll tax to maintain pay-as-you-go financing ofthe program. The OASI payroll tax rate remains at10.6 percent through 2024, rises to 13.7 percentduring 2025–2029, moves to 14.6 percent for thetwo decades between 2030 and 2049, rises to16.0 percent for 2050–2059, and then remains at16.4 percent beginning in 2060. The long-term risefrom 10.6 percent to 16.4 percent represents a taxincrease of nearly 55 percent.

Scheduled tax increases have been designedunder single-scenario assumptions that combinethe Trustees’ Report’s intermediate-cost assump-tions with the Census Bureau’s mid-range mortal-ity assumption. In that scenario, the tax scheduleproduces a near zero estimate for both the long-runactuarial deficit and the actuarial deficit in 2070.As shown in table 3.5, in results that explicitlyrecognize demographic and economic uncertainty,these higher tax rates have only about a 25 percentchance of financing the program’s long-run costs.On average, the long-run deficit will be about two-thirds of 1 percent, but it has a substantial chanceof exceeding 1 percent. Variability in the 2070actuarial deficit is even larger, as would be ex-pected.

Selected cohort effects of the CBIT reform areshown in table 3.6. The two lifetime measures ofprogram payback—the ratio of present value ofbenefits to the present value of contributions andthe internal rate of return (the present valuediscount rate that equates the present value ofbenefits and contributions)—indicate that the 1976birth cohort fares relatively well. But rememberthat much of the ultimate rise in the tax rates isscheduled for the years after this cohort retires.Younger birth cohorts would have lower paybackratios and rates of return. Replacement rates varylittle across the scenarios. The low benefit avoid-ance rate—the fraction of retirement years duringwhich the average benefit across all male cohortmembers falls above $5,000 per year in 1993dollars (or $417 per month)—is 100 percent.

Reduced Benefits and Current Taxes (RBCT)

RBCT generic reform maintains current-law OASItax rates and gradually decreases OASI benefits tomaintain pay-as-you-go financing of the program.


54

Label Mean 10th Percentile 25th Percentile 75th Percentile 90th Percentile

Long-Run Cost Rate for DB Benefits plus DC Contributions (% of taxable payroll)

CBITa 14.61 12.81 13.61 15.58 16.41RBCTb 11.85 10.72 11.18 12.43 13.02IARBc 13.92 12.86 13.42 14.5 14.93

Long-Run Cost Rate for DB Benefits only (% of taxable payroll)

CBITa 14.61 12.81 13.61 15.58 16.41RBCTb 11.85 10.72 11.18 12.43 13.02IARBc 9.32 8.25 8.82 9.85 10.29

Long-Run Actuarial Deficit for DB Benefits only (% of taxable payroll)

CBITa 0.67 –0.66 –0.06 1.34 2.01RBCTb 0.51 –0.61 –0.16 1.09 1.65IARBc –0.23 –1.04 –0.7 0.19 0.6

2070 Actuarial Deficit for DB Benefits only (as % of taxable payroll)

CBITa 1.25 –2.32 –0.81 2.88 5.46RBCTb 2.11 –0.46 0.63 3.29 5.15IARBc 0.26 –0.89 –0.42 0.79 1.65

Table 3.5Effects of Generic Reforms on OASI Program and on Economy

Note: See text for complete description of reforms. Results are based on 1,000 stochastic scenario values. Actuarial deficit isnegative of Trustees’ Report’s actuarial balance; long-run refers to Trustees’ Report’s summarized rate, which is calculated overthe 75 years ending in 2070.aCBIT: current-law benefits and increased taxes to maintain pay-as-you-go financing of pure DB Social Security program.bRBCT: reduced benefits and current-law taxes to maintain pay-as-you-go financing of pure DB Social Security program.cIARB: introduce 5 percent personal retirement accounts and reduce DB benefits gradually and DB payroll tax by 5 percent after

40 years of paying transition cost to a more fully funded mixed DC-DB Social Security program.

The decline in benefits is represented in this reformby a gradual reduction in the generosity of theprimary insurance amount (PIA) formula over theyears from 1999 through 2022. The magnitude ofthe across-the-board reductions in 2022 leavesbenefit levels 28 percent below that called for bycurrent-law benefit policy. Current-law cost-of-living-adjustment policy has not been changed inthis reform.

Scheduled benefit reductions have beendesigned under single-scenario assumptions thatcombine the Trustees’ Report’s intermediate-costassumptions with the Census Bureau’s mid-rangemortality assumption. In that scenario, the benefit-reduction schedule produces a near zero estimatefor both the long-run actuarial deficit and theactuarial deficit in 2070. As shown in table 3.5, inresults that explicitly recognize demographic andeconomic uncertainty, these lower benefits have

only about a 30 percent chance of financing theprogram’s long-run costs. On average, the long-rundeficit will be about one-half of 1 percent, but it hasa substantial chance of exceeding 1 percent.Variability in the 2070 actuarial deficit is about thesame, but it averages slightly more than 2 percent.This is significant given the lower cost of thescaled-back program.

Selected cohort effects of the RBCT reform areshown in table 3.6. The two lifetime paybackmeasures are somewhat below the correspondingCBIT reform estimates. This is because the fulleffect of the benefit reductions are experienced bythe 1976 birth cohort when they retire in the early2040s. The lower replacement rates vary littleacross the scenarios. The lower benefits are stillhigh enough to keep cohort average benefits atlevels that produce almost complete low-benefitavoidance.

55

Chapter 3

Table 3.6Effects of Generic Reforms on 1976 Birth Cohort

Label Mean 10th Percentile 25th Percentile 75th Percentile 90th Percentile

Lifetime Nominal IRR on benefits and contributions for men (%)

CBITa 5.58 3.77 4.58 6.57 7.43RBCTb 4.87 3.06 3.88 5.83 6.70IARBc 4.90 2.78 3.60 6.00 7.36IARB-aad 4.67 2.65 3.46 5.73 6.90IARB-nie 4.81 2.72 3.63 5.83 7.17IARB-aanif 4.57 2.61 3.48 5.49 6.74

Low Benefit Avoidance Rate for men (% of retirement years)

CBITa 100 100 100 100 100RBCTb 100 100 100 100 100IARBc 84.43 81.37 82.72 85.55 87.16IARB-aad 84.37 81.34 82.64 85.51 87.13IARB-nie 84.39 81.42 82.71 85.53 86.97IARB-aanif 84.32 81.39 82.64 85.46 87.00

Lifetime Ratio of PV benefits to PV contributions for men


Replacement Rate for steady average male earner (%)


Note: See text for complete description of reforms. Results are based on 1,000 stochastic scenario values. Members of thisbirth cohort are now 20 years old and will be age 65 in 2041.aCBIT: current-law benefits and increased taxes to maintain pay-as-you-go financing of pure DB Social Security program.bRBCT: reduced benefits and current-law taxes to maintain pay-as-you-go financing of pure DB Social Security program.cIARB: introduce 5 percent personal retirement accounts and reduce DB benefits gradually and DB payroll tax by 5 percent

after 40 years of paying transition cost to a more fully funded mixed DC-DB Social Security program. The basic IARB reformassumes life-cycle asset-allocation behavior and complete conversion of account balances into indexed annuities atretirement.

dThe IARB-aa variant is the same as IARB except that a 40/60 equity/bond asset allocation is assumed at all ages.eThe IARB-ni variant is the same as IARB except that account balances are converted into non-indexed (rather than indexed)

annuities at a lower price.fThe IARB-aani variant combines the two behavioral assumption changes of IARB-aa and IARB-ni.

balance using a life-cycle asset-allocation strategythat calls for investing completely in equities whenyoung and for the equity fraction to decline gradu-ally to 23 percent beyond age 60, with the bondfraction rising. Individuals are also assumed at

DC Accounts and Reduced DB Benefits (IARB)

IARB generic reform introduces a 5-percent-contribution personal retirement account in 1998.Individuals are assumed to invest their account


56

retirement to convert all of their account balanceinto an indexed annuity, which is priced assuminga continuation of recent mortality decline rates andusing a real rate of interest calculated with anexpected rate of inflation that is a moving averageof recent inflation rates.

Also, as part of the IARB reform, the definedbenefit OAI program (payroll-tax financed OASIbenefits first received by those age 60 or older) isgradually scaled back. The benefits are scheduledto decline gradually from 1999 to 2040, when initialOAI benefits would be reduced 70 percent belowcurrent-law levels. The OASI payroll tax, which iscurrently scheduled to be 10.6 percent during thenext century, is scheduled to decline—but notgradually like the benefits. The combination of5 percent account contribution rate and reducedOASI payroll tax rate would be 10.6 percentbeginning in 2040. But during the first four decadesof the next century, the 10.6 payroll tax rate wouldremain in place to finance the cost of transitionfrom pay-as-you-go financing to more fully fundedfinancing. During these four decades, the combinedtax/contribution rate would be 15.6 percent; after2040, the combined rate would fall back to10.6 percent.

This simple scheme for paying the structuralreform’s transition cost means that those cohortsworking during the first four decades of the nextcentury will bear the cost of the transition. Inparticular, the 1976 birth cohort, whose memberswill be age 65 in 2041, will be one of the cohortsmost burdened by the transition costs. Youngercohorts will experience significantly higher paybackmeasures because their members will have moreyears of work after 2040, when the combined tax/contribution rate falls back to 10.6 percent.

As shown in table 3.5, in results that explic-itly recognize demographic and economic uncer-tainty, the scaled-back DB portion of OASI experi-

ences relatively little variability in its long-run and2070 actuarial deficit. The 5.6 percent payroll taxrate that is in effect after 2040, when the gradualbenefit reductions have been completed, is suffi-cient to finance the pay-as-you-go DB portion ofOASI with relatively little variability in the 2070actuarial deficit.

Selected effects of the IARB reform on the1976 birth cohort are shown in table 3.6. The twolifetime payback measures are about the same asthose for the RBCT reform and somewhat below theCBIT reform. It is surprising that the IARBpayback ratios are as high as they are because thiscohort bears the heaviest burden of financing thereform’s transition costs. That the modest paybackmeasures are caused by the transition cost taxburden can be seen in the fact that the averagereplacement rate under the IARB is significantlyhigher than under either of the two nonstructuralreforms. As expected, the replacement rate with alarge DC component to Social Security experiencesmuch greater variability. There is also less thancomplete avoidance of low average benefit levels.

Cohort effects of three variants of the IARBreform are also shown in table 3.6. The nature ofthe alternative assumptions concerning individualaccount behavior is described on table 3.6. The age-inflexible 40/60 equity/bond asset-allocation variant(IARB-aa) has estimated cohort performancemeasures that are less desirable than those esti-mated for the life-cycle asset-allocation behaviorassumption (IARB). The mean payback ratio fallsfrom about 0.77 percent to 0.69 percent, and themean replacement rate declines from about58 percent to 53 percent. These differences arerooted in the fact that the life-cycle asset-allocationbehavior produces balance-weighted account ratesof return that average about 0.8 percent abovethose for the age-inflexible 40/60 equity/bond asset-allocation variant.

57

Chapter 3

4Comments on the EBRI-SSASIM2Model as of January 1997by Steve Goss

■ IntroductionThe Employee Benefit Research Institute’s (EBRI)Social Security reform analysis model—EBRI-SSASIM2—is extremely interesting; I am lookingforward to its completion for a number of reasons.Not the least of these is the number of importantobservations and findings that will result fromstudying this topic. Already, a few items haveemerged from the development of this model thatwill be very useful to the work we do at the Office ofthe Actuary at the Social Security Administration(SSA). As mentioned, they have selected 13 or14 key variables with means based on the interme-diate assumptions in the Social Security’s Trustees’Report,1 and the standard deviations are set basedon high-cost and low-cost assumptions to developthe stochastic model.

■ Examining the NumbersOne thing that struck me immediately in theresults occurred with runs 201, 202, and 203. Thesemodel runs are not stochastic runs; they are whatwe might refer to as “deterministic” runs. They arecontrasted with the trustees’ alternatives I, II, andIII, or the low-, intermediate-, and high-costalternatives, respectively. The results are verysimilar, but they are not symmetric. The high-costnumber is much farther from the intermediate-costnumber than is the low-cost number. In contrast,the distributions of the long-run average cost rate,and the long-run actuarial deficit that came out ofrun 205, which is the 1,000 simulation run, are

very symmetric, unlike the trustees’ three basicalternatives.

I assume that this result is related to the factthat the mean long-range actuarial deficit for theOld-Age and Survivors Insurance (OASI) programin the stochastic run 205 is 2.5 percent, while theintermediate assumption of the Trustees’ Reportprovides a 1.85 percent long-range actuarial deficit.There is quite a large difference between these twonumbers, and this difference is apparent before thechange in mortality assumption to the intermediatelevel used by the U.S. Bureau of the Census.

I had developed an explanation for thisdiscrepancy that seemed reasonable until I lookedat the graphs related to the stochastic run 205,which include the distributions of the cost rate andthe one-year actuarial deficit for the year 2070.These distributions are not symmetric. In fact, inthese distributions, the values that are in theTrustees’ Report for the intermediate projection areequivalent to the median, or the mode, the pointwhere the highest level is reached on these fre-quency curves. That is not the case for the long-range average cost or actuarial balance in run 205.This indicates that there are possible inconsisten-cies in these relationships that need further study.

One conceivable explanation is that when wedefine the intermediate assumptions for theTrustees’ Report and set these as the mean for eachparameter in the kind of stochastic approachdeveloped by Martin Holmer,2 it might result in anaverage cost or actuarial deficit that is, in fact,different from the intermediate. That seems to bewhat is happening. However, this explanation

1 Board of Trustees, 1996 Annual Report of the Boardof Trustees of the Federal Old-Age and SurvivorsInsurance and Disability Insurance Trust Funds(Washington, DC: Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability InsuranceTrust Funds, 1996).

2 See Martin R. Holmer, “EBRI Social SecurityReform Analysis Project Progress Report: Phases 1and 2,” in this volume.


58

appears to fail, leaving questions aboutprecisely what is occurring because we are notgetting that same result for the 2070 values. Thereappears to be an inconsistency.

■ Mortality Improvement andFertility Rates

Mortality improvement is one of the integralassumptions that the Trustees develop each yearfor the Trustees’ Report. A few years ago, theBureau of Census developed mortality projectionsdifferent from those used at the SSA. We at theOffice of the Actuary believe that the assumptionsused for the Trustees’ Reports are reasonable andappropriate. I would point to a couple of items toexplain our reasoning.

First, the numbers discussed by MartinHomer3 indicate that if the intermediate Censusnumbers were used, the average stochastic long-range deficit for the Old-Age and Survivors Insur-ance (OASI) program would increase by about0.7 percent of payroll, from about 2.5 percent up to3.2 percent. The Trustees’ Report provides sensitiv-ity analysis using high-cost mortality assumptionswith everything else on the intermediate basis. Inthis case we get about a 0.8 percent differential forthe Old-Age, Survivors and Disability Insurance(OASDI) program. So there is attention to uncer-tainty and variation in the Trustees’ Report.However, there is an important observation to bemade here. The Advisory Council’s Technical Panel,where the stochastic approach under discussionwas developed as it relates to Social Security, alsoconsidered mortality.

The panel’s best guess about future mortalitywould be to follow the approach used by Ronald Leeand Lawrence Carter in their September 1992article in the Journal of the American StatisticalAssociation. When we used that approach as anassumption for the ultimate rates of improvementin mortality, it resulted in a worsening of long-range OASDI actuarial deficit by 0.25 percent ofpayroll. This was smaller than we had expected.The approach used by Lee and Carter essentiallywould take the average annual rates of improve-ment in mortality by age and sex so far this century

and extrapolate them into the future. When that isdone, the largest differences from what we areusing in the Trustees’ Report occur at young agesbefore entrance into the work force. Thus, when weapplied the Lee and Carter method for mortalityrate improvement at these very young ages, we endup with very large rates of improvement. You mightthen ask, “What is the implication for SocialSecurity funding of having very rapid improvementin mortality at young ages?”

As far as Social Security is concerned, veryrapid improvement in mortality for the young isequivalent to having more births in the population.The reason is that more people will survive up tothe age at which they will be contributing to thesystem and having children. It ends up having apositive effect. That is why we had only about anegative 0.25 percent overall effect on the actuarialbalance as a result of testing the Advisory CouncilTechnical Panel’s approach to mortality.

The Technical Panel suggested anotherdemographic change, that of raising the fertilityrate. Purely by coincidence, the magnitude of theeffect of that change happens to be the same as theeffect of their suggestion on mortality—but in theopposite direction. Consequently, the two canceledeach other in terms of the effect on the overallactuarial balance.

That, however, is not the end of the story.Lower mortality rates and higher fertility rateswould have implications for some of the policyquantitative measures with which we have beenworking and that the model presents. As discussed,putting together the Technical Panel’s ideas aboutmortality and fertility would have essentially noeffect on the long-range OASDI actuarial deficit;however, it would increase the money’s worth ratioand the internal rates of return on a defined benefitplan. Of course, the reason for this is that longerlife expectancy under a defined benefit plan resultsin receiving the benefit for more years.

The other observation is that the replacementrate would of course be lower under the definedcontribution plans as a result of using the assump-tions of the Technical Panel. Longer life expectancymeans that if you have accumulated a certainamount of money in the defined contributionaccount as of retirement age, you will have tospread it over more years. So the replacement rateswould drop.3 Ibid.

59

Chapter 3

■ Effects of SimplificationThere are a number of caveats in terms of thepresent-law simulation developed by the EBRImodel. Some relate to the compromises and simpli-fications that are required to develop a new model.Another occurs in the development of the averagebenefit level. As you project into the future, understeady circumstances you can roughly assume thatthe benefit level for the current Social Securitysystem has a specific relationship to the averageearnings level, particularly the average earningslevel at the point in time when a worker retires.That is essentially the way that EBRI’s model isoperating at this point. However, there are somepotential problems with that simplification becauseone of the stochastic variables is the participationrate in the labor force.

For example, if in one of your runs theparticipation rate in the labor force turns out to besignificantly lower than it is under the average run,workers will have more “holes” in their earnings’histories. Therefore, the relationship between theirfinal earnings level and their benefit level would bedifferent from that of the average case. Theiraverage earnings over their lifetimes will be lowerrelative to the final earnings level. This is anexample of the kind of relationships you discoverwhen you develop any model, and it also invitesfurther discussion of what needs to be studied anddeveloped.

■ Developing Policy AlternativesAs you consider the alternative solutions to thelong-range OASDI financing problem presented inthis report, remember that the “average” deficitthat is being generated under this stochastic modelis 2.5 percent of payroll, versus the 1.85 percentlong-range deficit for the OASI program in theTrustees’ Report. This clearly would have implica-tions for the design of your solution if you weredesigning your solution based on this model. Note,also, that the model as presented pertains only tothe OASI program. Including disability is crucial.The Advisory Council plans, for example, and manyother plans under discussion, have very differenteffects on Disability Insurance (DI) benefits. It isessential to have a whole range of benefits that areconsidered under the OASDI plan.

■ Macroeconomic ImplicationsThe macroeconomic component planned for theEBRI model is not operational at this point, andtherefore we cannot comment on it. Many othermodels, including the model that we use at theSSA, do include some built-in macroeconomicfeedback, particularly when we model changes inthe normal retirement age and related proposals.Our structure models behavioral changes in thelabor force, which would feed through into theeconomy. Many other macroeconomic effects thatcould be expected to have impact are not includedat this point. This presents a real challenge for themodeling community as a whole.

■ Policy Performance MeasuresThe policy performance measures described in thisreport are familiar to many of us who have beenworking in this area. As described, there are twobasic measures. One is related to the average of allsingle workers; the other is related to the stylizedindividual. The stylized individuals are moresimilar to the hypothetical workers for which wedevelop analysis at the SSA. Stylized individualsare those for whom you specify an earnings leveland a marital status.

Under the stochastic model discussed, youmust specify three things for the stylized individu-als. One is the relative earnings level for theworkers over their career lifetime; another is thespecific age at retirement; and the third is age atdeath. Specifying the age at death is a very inter-esting adaptation of what has been done in thepast. In our work and in the work of most otherentities that have used money’s worth ratios, age atdeath is specified in a more distributional way. Thisallows the age at death to vary. Here it actuallywould be specified. Therefore, these stylizedindividuals end up answering different questionsthan those asked in the analysis for Social Securityhypothetical workers.

When you allow age at death to vary acrossthe whole possible distribution, you end up with theexpected money’s worth ratio or the expected rateof return. Under this stochastic model, if youactually specify age at death, you get the rate ofreturn and money’s worth ratio assuming thatspecific age at death. We have not pursued that


60

specification because you cannot anticipatethe age at death. We have assumed that mostpeople, not knowing even the average, or expected,age at death, would prefer analysis on a distribu-tional basis. However, this “stylized” approach doesanswer a different question than has been ad-dressed previously, and I believe there is real valuein that discussion.

The policy measures for the stylized individu-als are based only on single workers, and, at thispoint they exclude disability and survivor benefits.With respect to money’s worth ratios and internalrates of return, it is important to make furtherchanges as this model progresses. It will be veryimportant to ensure that the disability and survi-vor components of benefits under the system areincluded. In addition, the model’s stylized individu-als have a salary scale, as opposed to a flat wage-indexed earnings level. That will make a differencewhen comparing policy alternatives and provides agenuine contribution to the discussion.

■ ConclusionIn conclusion, this model provides positive informa-tion that will be useful in the further developmentof our work. Yet there are some questions that pointto the need for further work to develop this model.

One is the specification of specific legislativeoptions. These must be precise if you are going totry to use this model to determine the way aparticular option would operate relative to currentlaw. There may not be as much precision in theoption specification of this model as we would likeat this point. Attention also should be paid to theDI and survivor benefits to make sure those arefully represented. There also are some questions onsome of the policy variables about single-onlyworkers versus married workers and those withchildren.

My final observation concerns the expectationthat the completed stochastic model will be opera-tional on any desktop personal computer and thatanalysis would be attained by punching a few keysto select a few parameters. I do not know if thatwill be the result. My suspicion is that the detailrequired to get as far as Martin Holmer has inmodeling—along with the detail we have experi-enced in modeling at the SSA—does not suggestthat. Another problem occurs when you try tomodel new options that arise from time to time.These new options often require going back andreprogramming because we cannot anticipate everypossible model option. I suspect that the evolutionof the EBRI model will be similar.

61

Chapter 3

Social Security Reform: Beyond theModelsby Marilyn Moon

■ IntroductionI want to step back and take a broad view of thequestion of how a model like the Employee BenefitResearch Institute’s (EBRI) SSASIM2 PolicySimulation Model can play a role in policy discus-sion. It is important to understand the limitationsof these models. Ultimately, what we want to do isfind ways to understand the implications of variousoptions and then to explain them to people in waysthat are meaningful to a broad policy audience.Sometimes models turn out to be the tools thatanalysts use in making broad generalizationsbeyond what is justified by the analysis. Thus it isimportant not to get so enticed by the apparentobjectivity of modeling exercises that we forget thatsomeone is interpreting the results.

We have come to a very important point in theanalysis of public policy surrounding Social Secu-rity in which the complexity of the proposals oftenoutstrips our ability to carefully analyze them.EBRI’s model is seeking to expand our ability tolook at new issues. In particular, data on theuncertainty and the risk involved in proposals withvarying amounts of “privatization” should make areal contribution. Differences in risk represent acrucial aspect of understanding differences betweenplans that invest in the stock market and thetraditional Social Security program. But all modelsface many challenges in this complex environment.

First, incremental changes are easier todiscuss. By considering one aspect at a time, forexample, it is easier to see potential impacts.However, many of the new Social Security reformproposals now being discussed would change manyelements at once. This makes it difficult to disen-tangle various elements and understand theirseparate implications. What are the separateimpacts of establishing private accounts in theSocial Security system and of changes in the degree

5of redistribution that will occur across workers?These issues are clearly interrelated, although theyraise different points and they can be combined indifferent ways across proposals. This introduces theadditional question of how, then, to find ways to usemodels to separate out these different dimensions.

It is also difficult to find ways to talk aboutconcepts that Americans don’t understand well,whether or not they are included in models. Inparticular, risk is a difficult concept. The trade-offbetween higher returns and higher risks is notgrasped by many. Let me illustrate with an anec-dote in terms of my own family. My sisters-in-law,anticipating an inheritance down the way, are veryconcerned about my mother-in-law’s choice of amanager for her funds. Since a bank is managingher assets, they think that the bank can guaranteethe principal and get a guaranteed 15 percent to20 percent return. They don’t understand that itmight not work that way over time. Their expecta-tions have been shaped by the current market.

It is difficult to explain to people, when theyhear the magic words “high returns”—particularlywhen the Dow is around 7000—that there isuncertainty that such high returns will always bethere or that everyone will get them.

The challenge then is to find new ways toanalyze and interpret findings as well as to run themodels. My caution today is that we not be toocharmed with the elegance of the models and weopt for utilitarian approaches whenever possible.That means you have to make very tough choicesabout which things you want to add to a model andwhich new factors you build in.

■ Assumptions and ProjectionsAs a first step, it’s very important that we under-stand the current system and the future projec


62

tions. As Martin Holmer1 said, the assumptionsused in the Social Security trustees’ report serve asthe basis for this model, with a few exceptions.These assumptions matter, and they are verydifficult to develop. Setting them each year isclearly the source of a great deal of angst and hand-wringing by the trustees and the actuaries.

I would only add that it is instructive to lookback 75 years in considering how difficult it is tomake future projections. Should we use all thatpast experience in projecting forward? Does recentexperience count more than experience in 1925?For example, productivity has been quite low in thelast 20 years, while a longer look back would givehigher average productivity figures. How stablewill the future be?

In the case of mortality assumptions, it is nowfashionable among demographers to look back intime over a full 75 years and then project the sameexperience forward. This results in an assumptionthat women will continue to outpace men in termsof their life expectancy indefinitely, although recentexperience indicates that this is not what is hap-pening. These are just two pieces of the puzzlewhen you are trying to figure out what to say aboutwill happen 75 years into the future.

It is important for all of us to understand thatprojections and modeling the future are interestingand important exercises, but these are exercises.We all know—those of us who engage in this— thatwe are going to be wrong. We just don’t know howwe are going to be wrong.

■ Issues to ConsiderNow let me mention a number of issues that shouldbe included in the discussion about options for

Social Security reform, only some of which can bemodeled. First, would splitting the Social Securityprogram into a system of private accounts and abasic fundamental guarantee be viewed as solvingthe problem? Will it help raise additional nationalsavings and encourage people to save more forretirement?

A second and related issue is whether it willbe irresistible, under a privatized system, to allowpeople to dip into their personal accounts for otherreasons such as buying homes or paying for educa-tion and medical expenses. These are very popularproposals now with individual retirement accounts(IRAs), and this will also have implications interms of long-run modeling of various options.

Third, how will we deal with vulnerablepopulations and redistribution? The issue of thetreatment of dependents and survivors is critical toanswer this question and one that should be addedto the model. But the debate that will occur overprivatization will sometimes be dominated by theargument that, “This is my account and it is mymoney and I’m not sharing it,” as opposed to othermore objective measures of the advantages anddisadvantages. This is an element of the debatethat will not be addressed at all by the models. Thisis a very subjective and important judgment that isgoing to be made in the policy discussion.

■ ConclusionIn sum, there is a whole range of crucial questionsthat models, no matter how sophisticated, cannotanswer. We need to find the right balance betweena sophisticated model that we think addresses theright issues and the crucial decisions that peoplewill need to make without benefit of models. Wehave to be careful about the arrogance of theinfallibility of models when we know that somevital questions are being left out.

1 See Martin R. Holmer, “EBRI Social Security ReformAnalysis Project Progress Report: Phases 1 and 2,” inthis volume.

63

Chapter 3

6The Importance of SystematicAnalysisby William Beeman

■ IntroductionIn working on a policy statement about SocialSecurity reform, the trustees of the Committee forEconomic Development (CED) identified two majorissues that they believe must be addressed: thefiscal imbalance of, and the declining return on,Social Security contributions. When you look atthese two issues, you discover that a great deal ofsystematic analysis is required to merely describethem; even more sophisticated analysis is neces-sary to prescribe policies to deal with them. Forthat reason, the Employee Benefit ResearchInstitute’s (EBRI) sponsorship of the work of thePolicy Simulation Group is important. The modelsimulations may turn out to be wrong; almostcertainly they will be wrong. Their value, however,is that they force us to think systematically aboutpolicy options.

As one might expect, there is a great deal ofcareless talk about policy options; I run into itconstantly. For example, when people discoveredthat CED was going to work on Social Security,many of them came to my office to lobby for aparticular viewpoint. When I asked them funda-mental questions about their position on reform, Idiscovered that many had not subjected their planto any systematic analysis. The extreme caseoccurred last fall when I asked an advocate ofprivatization the fundamental question, “What doyou propose with regard to transition costs in yourproposal?” I was amazed that he did not knowabout the concept of transition costs arising fromprivatization. I spoke to others who were verysophisticated about the concepts, but they have areal need for this kind of analytic model to helpthem quantify their proposals. The projections,produced both by the Social Security Administra-tion (SSA), including the information compiled forthe Advisory Council, and, I presume, eventually bythis model, and their policy analysis are fundamen-

tal to any kind of serious evaluation of majorchanges in policy.

■ Examining the BaselineProjections

Originally, I intended to discuss the model itself,but the details of the model are not available yet. Itis hard to evaluate a model on the basis of a fewsimulations. Nevertheless, there are a number ofinteresting aspects of the model’s baseline projec-tions. For example, a lot of people talk about theproblems of Social Security being far in the future.But if the model simulations are correct, thoseturning age 65 when the trust funds are depletedare now about 35 years old. And looking at therange of probabilities, they might be 40 years old atthe moment. The point is that the fiscal imbalanceis not some abstract thing that should be of nointerest to most people. For those who find theprojections too abstract, I like to place them in amore personal context. I note, for example, that mydaughter will be 56 years old when the funds aredepleted and my grandson will be 30. My “cohorts”will be dead. Actually, a few may survive (I wouldhave to live to age 95), but most of those will havefew remaining assets and will be totally dependentupon Social Security at a time when, at leastaccording to the projections, the SSA revenueswould finance only about 70 percent of benefits.

■ Handling Risk and UncertaintyThe most interesting aspect of these model simula-tions, the one I find most intriguing, is the treat-ment of risk and uncertainty. It forces us to thinkabout the range of likely outcomes and aboutdesigning reforms that can accommodate unex-pected developments. CED’s analysis of policyoptions, published in February 1996, was deeply


64

influenced by the level of uncertainty. Unfortu-nately, I did not have Monte Carlo simulations topresent to our trustees, but we discussed uncer-tainty a great deal, and the design of reformsproposed by CED reflects this uncertainty.

The reforms proposed by CED handle uncer-tainty about future revenues and costs in two ways:First, the statement endorses change that wouldpermit the system to compensate automatically forprojection errors; second, it proposes changes(mostly a reduction in the growth of benefits) thatwould do more than merely eliminate the projectedactuarial deficiency. In this way, if the projectionsturn out to be correct, it would be possible to haltthe phase-in of further changes in the system. IfCED’s proposals were adopted, it would be unlikelythat we would need to address the fiscal imbalancein Social Security in another 13 years.

When completed, the Monte Carlo experi-ments should be very useful in the debate about thedesign of reform; however, in the past such analysishas prompted me to consider issues such as whathappens if events fall outside these ranges. Forexample, a news magazine recently reported thatscientists are becoming very interested in thefundamental determinants of life span. One storyannounced that scientists now know how to in-crease the life span of fruit flies by 50 percent andthat they are extending their studies to other kindsof creatures. If this work has implications forhumans, we will need a different type of recommen-dation for the SSA. Perhaps it should pay thesescientists to drop their research. Alternatively, thetrustees could ask the scientists to research thefollowing question: How do you induce fruit flies todelay retirement? Or perhaps they can identify thegene that makes rats stay in the rat race longer.

■ ConclusionsI urge the model builders to continue their workand to progress to the macroeconomic issuesbecause, in my judgment, one of the most impor-

tant criteria for evaluating reform proposals is theeffect on national saving. Speaking only for myself,I would not endorse a reform proposal that did notat least have a very good chance of significantlyraising national saving.

I also believe that one should proceed withgreat care in using data both from these modelsand from calculations by the SSA. For example, theEBRI model discusses money’s worth ratios bycohorts. The pay-as-you-go option predicts thatyoung workers should worry a great deal aboutSocial Security, and they are worrying about it. Andit says that the outlook is worse for their children.As you know, with the pay-as-you-go option, taxincreases are postponed until they are necessary; ineffect, the tax burden is passed on to young work-ers, creating lower returns for each new generation.

The rate of return analysis is another criticalconsideration. On the other hand, is it reallyvaluable to compare reform proposals using thisanalysis? Suppose you wanted to create a reformproposal that showed a very high rate of return.What would you do? The answer is (1) look forfinancing outside the Social Security system, or(2) delay tax increases as long as possible. If onewere an extremist on the return issue, I would say,“Well, lower the payroll tax and increase thefederal deficit to pay for it until we get the rate ofreturn on contributions rising in the projections.” Itis important to remember the sources of differencesin these projections. Some plans, for example, dohave deficits; they finance the transition throughdeficits. Such proposals are likely to increase therate of return within Social Security and decreaseit for the economy as a whole. So, it requires agreat deal of skill and thought about the use of thevarious kinds of analysis provided by the simula-tions, but they are absolutely essential. I stronglyencourage EBRI to continue to support this projectand I am looking forward to the macroeconomicsimulations, which I view as the most importantaspect of the modeling.

65

Chapter 3

7■ IntroductionA few years ago, a couple of us at Brookings wrote abook about the cost and affordability of SocialSecurity.1 The public trustees of the Social SecurityAdministration (SSA) commissioned our study, andwe tried to be very evenhanded and scientific. Wewrote soberly about the unpleasant policy choicesfacing the country, and we tried to show whycertain policy choices might be preferable to thestatus quo. It would not be exactly accurate to saythat important policymakers ever read our book.But aides to important policymakers—or perhapsthe unpaid summer interns of the aides—may havedone so. Of course, our book has not had a notice-able impact on public policy ... so far. The Presidentand Congress have not followed our excellentadvice, and perhaps they never will.

The title of the book, Can America Afford toGrow Old?, raised the question under discussion. Ithought the first 20 pages of the book clearlyconveyed the answer as, “Yes, we can afford to growold.” But more than half the news articles writtenabout the book concluded that the answer was, “No,there’s no hope. We can’t afford to grow old.” Thismisinterpretation of our message may reflect acommon problem among journalists, who mayreason that they never will see their names in printif they write a story about good, or, at least, am-biguous news.

Our message, which was fairly simple, saidthe opposite: “Look, the country’s population isgrowing older. This is clearly going to raise thepercentage of the population that draws SocialSecurity benefits in the future. But the resulting

Can America Afford to Grow Old?by Gary Burtless

growth in spending is manageable, about 2 percentto 3 percent of Gross Domestic Product for SocialSecurity.” The extra spending could be financed outof growing incomes. Moreover, the amount of extraspending actually could be scaled back if the nationundertook prudent reforms before the huge in-crease in the number of pensioners.

We even suggested a public policy to reducethe future burden of Social Security. In particular,we could save a larger Social Security surplus. Thatis, the country could accumulate Social Securitysurpluses at a faster rate than will occur undercurrent law in one of two ways—either by hikingthe payroll tax or reducing promised benefits in thenear future. At the same time, it could reduce oreliminate the deficit in other government programs.That combination of policies would increase govern-ment saving and, in turn, raise the nation’s capitalstock over the medium and the long terms. Ifnational investment and saving were higher, futurenational output could be higher. The nation couldbe paying for a larger amount of public spending onthe nation’s elderly, but it would pay for it out of alarger economic pie, leaving a bigger slice of thatpie for future workers to consume. Thus the burdenof supporting an aged population would be less-ened.

It is quite plain, seven years after publicationof our book, that this multidimensional policy is notabout to be adopted. Congress and the Presidenthave not taken any action to accelerate the growthin the Social Security surplus, although they havesharply reduced the deficit in the remainder of thefederal budget. Many believe that any attempt toexpand the Social Security surplus would be futileanyway. If policies were adopted to expand thesurplus, Congress and the President would simplyspend the bigger surplus on other governmentoperations, including Medicare.

1 Henry J. Aaron, Barry P. Bosworth, and GaryBurtless, Can America Afford to Grow Old? Paying forSocial Security (Washington, DC: The BrookingsInstitution, 1989).


66

actually reflect the recent experience of the UnitedStates. I do not think we can say that each of theassumptions of the Social Security trustees actuallyreflects recent experience in a symmetrical way.For a couple of assumptions, I think, the centralpoint estimate may be more optimistic than therecent experience. In addition, even the lowerbound estimate is more optimistic than I thinkmany people would accept for a lower boundestimate of what the future may hold. So what ifour modelers moved away from the trustees’assumptions and, instead, asked, “What has beenthe experience of the United States in the last25 years? Now let us bound that actual experiencewith symmetrical optimistic and pessimisticassumptions about the future.” That policy simula-tion would be of interest in addition to analysesthey already have performed.

Second, it is crucial that the modelers accountfor the feedback effects of these reform plans onsaving and investment and, hence, on futurenational income and future real wages. Thisextension is particularly important if we want toassess the plans that propose to increase nationalsaving, either through privatization or through anincrease in the size of the Social Security reserves.It is necessary to model the feedback effects toevaluate the primary advantage of increasingnational savings in advance of the baby boomgeneration’s retirement.

Finally, it will be useful for the authors toexamine several proposed actual plans. Theyshould evaluate not only the Advisory Council’sthree plans but also the plans suggested by othergroups. Some plans, for example, move even furthertoward full privatization of the nation’s retirementsystem.

■ The Questions for Generation XThe calculations in this paper lead to a conclusionalmost anyone would reach: Partial privatizationmay not be a good idea for workers in GenerationX. Under full or partial privatization, these peoplewill be asked to (a) contribute to a slimmed-down,defined benefit Social Security program; (b) contrib-ute to a new defined contribution private plan thatwill supplement reduced Social Security pensions;and (c) contribute a substantial percentage of theirearnings to pay off the accumulated liabilities of

■ Proposed ReformsSo this brings me to the recommendations of thelatest Social Security Administration’s AdvisoryCouncil. That council’s deep division was reflectedin the failure of any single proposal to gain amajority. All three plans to emerge from theCouncil tried to boost public or private saving,although they would attempt to accomplish this invery different ways. Thus, all three plans tried tosecure some of the potential gains that we couldenjoy by obtaining a higher return on the retire-ment system through investment in the capitalmarket. In the present environment of slow produc-tivity growth and slow growth in the labor force,returns in the capital market typically exceed thereturns that can be obtained in a pay-as-you-gosystem. All three plans would slow the future rateof growth of Social Security benefits. One wouldincrease the rate of growth of the Social Securityreserve. The other two would create a new pool ofprivate savings to finance mandatory definedcontribution plans.

■ The ModelThe Employee Benefit Research Institute’s (EBRI)simulation model—EBRI-SSASIM2—can help us tounderstand the three plans. This model, a worth-while effort that shows much promise, projectsinteresting results concerning the risk of adversedemographic and economic events. It demonstratesthe implications of different reform plans forpayroll contribution rates and cost rates and for thelong-term actuarial deficit as a percentage ofpayroll. It shows returns and money’s worthcalculations under various reform plans. It alsoexplores long-term replacement rates underalternative plans. This kind of information, as wellas information about the uncertainty of theseoutcomes, is necessary to consider how we shouldattempt to reform Social Security. Some extensionsof the model will be very helpful.

■ Obtaining SymmetryIn regard to examining risks, it would be useful ifthe modelers adopt symmetrical optimistic andpessimistic assumptions about the future. Theoptimistic and pessimistic assumptions should

67

Chapter 3

Social Security. These liabilities are owed to peoplewho are nearing retirement age or retirees whoalready are collecting pensions. In effect, Genera-tion X workers will be asked to pay for the retire-ment of two generations—their own and thegeneration immediately preceding them.

Let me add, however, that the burden can belightened in two different ways. This will becomeplain when the modelers account for the feedbackeffects of higher saving on the national economy.Faster growth of the economy, faster growth of thecapital stock, and faster growth of real wages wouldreduce the burden of old-age pensions on futuregenerations of workers. In addition, the burden onGeneration X would be lightened still further if theburden of paying for the accumulated liability werespread out over a longer time than 40 years. Oneidea is to impose a tax in perpetuity that pays theinterest on the additional debt that the countrymust accumulate to pay off the current accumu-lated liability of Social Security. But remember, ifwe pay only the interest on the accumulatedliability (rather than the accumulated liabilityitself) over the next 30 or 40 years, we will notenjoy a very big advantage from the point of view ofincreasing national saving. If we lighten the burdenon Generation X by extending the period overwhich we pay off the accumulated Social Securityliability, we will not boost national saving by verymuch over the next 30 or 40 years.

■ Cost-of-Living CalculationsFinally, let me address another important eventthat has implications for Social Security reform, therecent report of the Advisory Commission to Studythe Consumer Price Index.2 If the commission is

correct in believing the CPI overstates changes inthe cost of living by an average of about 1.1 percentannually, we have all received a very good piece ofnews. First, we are all richer than we thought wewere; compared with 40 years ago, we are55 percent richer than we thought we were. So weare already better off.

Second, in the future we will be richer thanwe thought we were going to be. If the trusteesthought real wages are going to rise by about40 percent between now and 2035, the commissionhas good news for them. It now turns out thatwages actually are going to rise 107 percent. Oneimplication of this is that if the payroll tax mustrise four percentage points to pay for the largerretired population in 2035, workers in 2035 will beearning 95 percent more than we do today, eventaking into account that they will pay a higherpayroll tax than we pay today. If this is true, I amnot going to worry too much about making toughchoices today. Why make consumption sacrificestoday that will help these workers trim theirpayroll tax by a couple of percentage points? Theywill be twice as rich as I am. Why should I scrimpand save on their behalf?

Third, payroll taxes are not going to have torise as fast as the trustees predict. Why? We cangive less money to future retirees than is predictedin the trustees’ report. We will not have to givethem cost-of-living increases that are as large asthe trustees assume. Instead, we might give themcost-of-living increases that average about1.1 percent less. I hope that in future implementa-tions of this model we will see the full implicationsof the commission’s report. I also hope, however,that no reader of this comment will believe I amseriously proposing a cut in cost-of-living adjust-ments amounting to 1.1 percent a year. To makesuch a proposal, I would first have to be persuadedthat the bias in the present consumer price index is1.1 percentage points. I have yet to be persuadedthat the bias is that large.

2 Advisory Commission to Study the Consumer PriceIndex, Toward a More Accurate Measure of the Cost ofLiving: Final Report (Washington, DC: Senate FinanceCommittee, December 1996).


68

69

Chapter 3

8■ IntroductionI want to make a few comments at the beginningabout the actual modeling exercise; in particular,about the mortality assumptions. In the presentversion of the model, the only thing that is reallyanalyzed is the effect of mortality. Most of the otherinteresting issues have not yet been included in thesimulation model. In particular, the crucial issuesof the effects on savings and economic growth,which I think ought to be the main criteria onwhich we decide whether reform is desirable, areexcluded.

■ Mortality AssumptionsThere are two issues of Social Security reformconcerning mortality. First, it is a very wise deci-sion to abandon the Social Security Trust Fund’smortality assumptions and move to the CensusBureau intermediate estimates. Second, theauthors also present a range of values throughthousands of iterations that indicates somethingabout the uncertainty involved with future mortal-ity. That is a useful exercise. However, I think wehave to be very cautious because the uncertainty ismuch wider, mostly in a pessimistic direction—pessimistic in the sense that we are going to livelonger. Therefore, projections for the Social SecurityTrust Fund may be more pessimistic than theseestimates indicate.

In an academic audience, the Trust Fund’smortality estimates would not be taken seriously asprojections of future mortality declines. Even theCensus Bureau estimates are viewed with somesuspicion as accurate estimates of what futuremortality decline estimates will be. Most of theestimates in the academic literature see muchsharper declines in morality, and, especially, thepotential for much larger fractions of the popula-tion who are old. If I had to take a point estimate, I

A Comment on Mortality Assump-tions, Low Wealth, and the Need forMore Savingby James P. Smith

would actually use the ones that Martin Holmer1

mentioned—the estimates that are being done byRon Lee at Berkeley with a series of his colleagues.

I would view Lee’s work as the best pointestimate of the future course of mortality. Toillustrate the implications: the number of peopleover age 85 according to the Trust Fund estimatesis 14 million; the Census intermediate estimate is18 million, and Ron Lee and company project21 million. I would not use the long-run costaverage to determine the implication of thesedifferent mortality rates. Instead, I would use the2070 numbers—that is the steady-state to whichwe are heading from different mortality estimates.The effects of using Lee’s estimates on the steadystate, 2070 numbers, in terms of the average cost ofmaintaining the current benefit structure, aremuch higher than the discussions at this policyforum have indicated. They would bring it up by atleast another two percentage points.

Lee’s estimates move the central tendency onmortality over to a higher number. We will have arange of possibilities outside that central tendencyof Holmer’s estimates—a different central tendency.Because of this, we have a different range of whatis possible than before. As a result, Holmer’soriginal range of estimates is not really an indica-tion of the real uncertainty if we do not get thecentral tendency right.

The seriousness of this problem is indicatedby the different estimates available from veryreputable scientists, whom I regard as experts onmortality (in particular, Ken Manton and JimVaupel). Instead of the 14 million projected by theTrust Fund, they are projecting 39 million peopleover age 85 in the year 2050. These numbers would

1 See Martin R. Holmer, “EBRI Social SecurityReform Analysis Project Progress Report: Phases 1and 2,” in this volume.


70

certainly not be my point estimate. But theyare much more plausible outcomes than the reverseon the other extreme, which would be no growth inthe population of people over age 85 compared tothe current estimates.

What we have then is a higher mean value formortality and a very skewed distribution, wherethe outcomes with much sharper declines inmortality more likely than the outcomes of rever-sals in the mortality process. All these consider-ations are leading to more pessimistic estimates ofthe cost implications and higher average costestimates for the Social Security program’ssustainability.

■ Mortality and Socio-EconomicStatus

The second problem with the way these mortalitynumbers are being generated is that there isconsiderable recent research showing that mortal-ity is very sensitive to measures of socio-economicstatus, such as income or wealth. Research on low-income individuals done by Lee Lillard indicatesthat if you take people in the bottom 10 percent andpeople in the top 10 percent, mortality differencesare sevenfold in terms of age-specific mortalityrates. Other scholars have estimates of a strongrelationship of mortality to wealth. This strongrelationship between mortality and measures ofsocio-economic status has implications for thedistributional equity of the program and also for itscost.

Even though Social Security is a progressivesystem in terms of benefits, people with higherincome still receive more benefits, and people withlower income receive fewer benefits. Decreasingmortality among those receiving more benefits, ifthe overall mean for mortality rates remainsconstant, could have a significant impact onprogram cost.

This relationship of socio-economic status tomortality also affects how we view Social Securityin terms of its equity. Once you introduce the factthat high-income people are going to live longerthan those with low income, about one-third of theSocial Security progressivity is eliminated by thisrelation.

■ Achieving Fiscal BalanceLet me make two other points. One of the thingsthat troubles me about this simulation exercise—not just this exercise, but all simulation models—isthat fiscal balance is achieved in one of two ways.We either reduce benefits or increase taxes. All theadjustments are self-contained within the SocialSecurity system. What we are leaving out, andwhat I think has to be included in the modelingexercise, is that these are not the only ways ofbalancing this system. In fact, this has not been theway we have been going about balancing thesystem in recent years.

What we have to add to this model is what thegovernment sector is going to do. If the tax rateneeded to sustain this program to the year 2070 isabout 20 percent for Social Security and another12 percent for Medicare, this is a combined tax rateof one-third on these programs. Therefore, we aremoving up toward a tax burden of 40 percent on anage-related transfer from everyone below a certainage to everyone else above a certain age. In time,that is the tax level to which the current level ofbenefits is pulling us. We have to ask ourselves, ifwe maintain the current structure of benefits, whatwill happen to the government sector?

I think what will happen is that other govern-ment programs will be severely reduced. This policyis the same as putting Social Security off the table,exactly what we are doing now. We are puttingSocial Security off the table and making all cuts inother social programs, many of which are hittinglow-income people very hard. If we take the posi-tion that we want to maintain an age-related set ofbenefits that imply a 40 percent average tax rate,we should include in the modeling exercise whatthe likely reaction is going to be in terms of the restof government’s expenditures. This policy does notseem progressive to me.

■ ConclusionI will conclude with some of my own recent re-search on the economic status of the elderly.

Using recent data funded by the NationalInstitute of Aging, we now have a much better ideaof the financial situation of the elderly and nearelderly. This financial situation covers all financial

71

Chapter 3

assets, broadly defined, including checking andsavings accounts and just about any financialassets you can think of. For people over age 70,those who are currently in the Social Securityprogram, financial wealth is very unequallydistributed and very small for most older house-holds.

The big news is not that the poor did not save.The news is that the average white household overage 70 has only $10,000 in total financial assets. Ifthey had to rely only on their past savings, it wouldget them by for about half a year. Even if youexamined people right before their retirement, youmight think they have a lot more on hand to dealwith the emergencies that might happen in theirlives. However, the average white household intheir 50s has financial assets of $17,000. That isequivalent to about half a year’s income. If they lostall other sources of support, they could generatehalf a year of income.

Any sensible and prudent Social Securityreform will reduce the level of age-related benefitsindefinitely into the future. However, unless peoplesimultaneously increase the level of their privatesavings, we are going to have many people in verydire straits indeed.

I will close with the following argument. Youcannot talk about Social Security reforms inisolation. You must also deal with encouragingsaving and economic growth. My proposal is, inaddition to changing the benefit structure of theseprograms, to introduce, immediately, a consump-tion tax so that we do not tax the creation ofwealth, either savings or investment. I would insistthat it be a progressive consumption tax. We knowthat most rich people favor a consumption tax notbecause it encourages growth, but because it isgoing to make them a lot of money. But we can havethe economic benefits of a consumption tax andmake it sufficiently progressive to eliminatelegitimate equity concerns.


72

73

Chapter 3

What Can We Learn from ModelingSocial Security?by Gene Steuerle

■ IntroductionHaving built a few models myself, I am well awarehow easy it is to comment on all of a model’slimitations. During an earlier discussion I had atthe Employee Benefit Research Institute (EBRI)about some of the model-building that we weredoing at Urban Institute, someone said, “I thoughtwe agreed about five years ago that some of thosedata are outdated and that we needed to move on tobetter behavioral assumptions.” All of us can be abit taken aback by comments on our models’limitations.

We are somewhat like Sisyphus, who aspunishment for having engaged in a bit of trickerywas required for the rest of his life to roll a ball upa hill. As soon as he got it near the top of the hill, itrolled back down again. He kept rolling it up and itwould continue to roll back down. That is what Ithink often happens when we try to build thesemodels. We think we’re going to roll this ball up thehill and come up with the ultimate model, and thensomeone comes along and says, “But you didn’t dealwith this,” and the ball goes back to the bottom ofthe hill.

■ Allocating ResourcesConcerning Social Security, I don’t think the issueis only one government program. I think we aredealing with a very broad societal issue of justwhere we want to go as a society and how we spendthe increments of our wealth as a society. I agreewith Gary Burtless1 that we can afford an increasein those payments to the aged that are defined asSocial Security. I think we can afford it if that iswhat we want to do. But we have essentially been

on a 50- or 60-year trend where tremendousportions of our increases in wealth have essentiallybeen paid to ourselves, for consumption in old age.

We have pursued this long-term trend notonly through Social Security but also throughprivate pension systems and through Medicare.This path of growth has been at an unsustainablerate; that is, at a rate that would eventuallyrequire us to spend more than 100 percent of ourincome on consumption in old age. We know wehave to move off that path. What makes it sodifficult is that, in addition, there is also thedemographic problem: a dramatic and rapid drop inthe ratio of workers to retirees. The baby boomers’retirement will accelerate this phenomenon.

Recent Congressional Budget Office (CBO)projections concerning Social Security, Medicare,and Medicaid indicate that an increase of abouteight to nine percentage points of Gross DomesticProduct (GDP) would be required simply to pay forthese programs. Two-thirds of that number repre-sents health care benefits, and I don’t think thathealth is something you necessarily want to leaveoff the plate when designing a reformed system. Ifwe translate CBO numbers to a tax rate, ourcurrent system already has built into it a tax rateof about 33 percent, from these government pro-grams alone. This tax rate has been built into theexisting system, and now the question is how can itbe cut back.

This 33 percent tax rate can be translatedanother way. Suppose we expect to live one-third ofour lives in retirement, we have a ratio of essen-tially two workers per one retiree, we expect tomaintain ourselves in retirement at about the sameincome level of income that we had before retire-ment, and we expect these retirement benefits to beprovided by other generations. Again, this requiresabout a 33 percent tax rate; i.e., these two workerswould have to pay about 33 percent of their income

1 See Gary Burtless, “Can America Afford to GrowOld?” in this volume.

9


74

each simply to support the third adult who isretired.

While Social Security by itself does not havethis type of replacement rate, once you startcounting spousal benefits, the fact that manybenefits are nontaxable, and the huge increase inhealth care costs—which are much higher amongthe aged—in fact, you do get closer to that type ofreplacement rate. I think this rate is largelyunsustainable, not because we cannot afford it butbecause it is an allocation of economic reserves thatdoes not make sense relative to other societalneeds.

■ Increasing SavingsThe specific plans advanced by the Social SecurityAdvisory Council only address a small part of thisproblem. They mention trying to find one percent-age point of GDP or, with luck, one and one-halfpercentage points of GDP, shifting the numbersaround one way or the other and hoping that thiswill have an impact on the much larger problem. Inthe so-called maintain benefits plan, the govern-ment reallocates its portfolio and shifts the assets itbuys—not necessarily increasing saving all thatmuch.

Another plan makes more of an attempt toincrease private saving, but one needs to be carefulin assessing how much. This plan tries to increasesaving by taking five percentage points out of thecurrent Social Security tax rate. That is about twoand one-half percentage points of GDP. Then itimmediately says, “Because there are so manyproblems with maintaining minimum benefits, weare going to borrow the equivalent of two percent-age points and have another tax increase of oneand one-half percentage points,” so the net changein tax rate ends up being only about one and one-half points. Again, this is in a system in whichmerely to achieve balance, the equivalent of a33 percent tax rate is promised for the future. Toget that rate down to 15 percent is an enormouschallenge.

■ The Work ForceI have some concerns with a lot of the modelbuilding surrounding “privatization”—even thoughI myself have favored some of the types of

privatization proposals. One real concern is that Idon’t think enough steel mills can be built to solvethe problem of the reduction in human capital. Thisquestion of building human capital into the modelsseems to me most difficult.

Let me put it in rather stark terms. If wewere to drop immediately today to the ratio ofworkers to retirees that we project will occur inabout 30 years, this would be equivalent to addingabout an additional 10 percentage points or more tothe unemployment rate. And only a miracle workercould manipulate one percentage point of a tax ratehere or there and address this type of problem.There will be a tremendous drop in human capital.Measured in terms of wealth, it is on the order oftrillions of dollars of decrease in our use of humanresources. One thing I hope that we start buildinginto our models is the effect of this decline inhuman capital and how we will cope with a muchsmaller labor force.

■ Retirement AgeOne difficult variable to model is the statutoryretirement age. Before I began to study the issue, Ihad always assumed that 65 was an appropriateretirement age, but I had no idea where the num-ber came from. It is embedded not only in ourpublic-sector laws but also in many private-sectorplans as well.

What this signal tells our society, in terms ofits expectations, is that we can expect be retired forone-third of our adult lives. What happens if youchange that signal? It could have enormous behav-ioral effects. I think that this is a vital, and prob-ably one of the most important, issues that face uswith respect to modeling changes in retirementprograms.

There is a closely related issue that I don’tknow how to model but consider crucial. Myimpression is that many executives in the privatesector are quite willing to take on the issues ofSocial Security but are unwilling to come to gripswith the issue of when to retire people within theirown private plans. They say, “We will work onSocial Security, but we still want to retire all thesepeople at age 55 and 60 because we are overpayingthem and cannot afford to pay them at currentlevels.”

75

Chapter 3

■ The Impact of GovernmentYet another modeling issue is how much govern-ment can affect net investment or saving. Thesimple model assumption that most people startwith, and then modify slightly, is that if govern-ment can somehow extract an additional dollar andput it in an account somewhere—be it a govern-ment account or a private account—it has increasedsaving by a dollar.

This might be true in terms of a singleindividual, but even there one needs to be careful.Investors in financial assets have also financed alot of borrowing on credit cards today. Secondarymortgages have also become very common. Theidea that a gross deposit increase of a dollartranslates into net saving increase of a dollar ismisleading. When one starts thinking aboutmodels, one has to ask about these shifts.

As I understand the EBRI model as it isdeveloping, it allows for some private-sector shifts.For instance, if one increases government-man-dated private saving or private deposits by a dollar,it might have some impact on private pensiondeposits. If I have a mandated savings plan andcontribute $1,000, the question is how much do Ithen put in my 401(k)? At that point, we are notdone because other shifts are also occurring.

Think of the market for loanable funds. Putanother dollar in deposits. Anyone can come in thatmarket and borrow, and many of the borrowers arenot taking that money and converting it to netinvestment. The crucial question then becomes,how much can government actually affect netsaving and investment, even if it can get an addi-tional dollar into an account and even if it builds upthe Social Security surplus?

A study I did for the Brookings Institutiondemonstrates that in a period when governmentwas saving much more, especially in the early post-World War Two period, there was a tremendousincrease in consumer borrowing at the same time,probably financed in part by the government’sefforts. A crucial issue that we need to be able tomodel is how much government can affect netinvestment and saving. We should model alterna-tive assumptions as well. One strong possibility isthat government, no matter how much it adds todeposits in one part of its system, will increase netsaving or investment only modestly.

■ Integrating Private Plans in theModels

This brings me to a related issue: the question ofintegrating private plans into these models. Wehave talked about mandating benefits, or mandat-ing savings in various accounts, but have notdiscussed how this integrates with private plansand private markets. This goes beyond the issue Iraised previously: If I have another dollar ofsavings in an account because of a government-sponsored mandate, how much have I affectedsaving in a 401(k) plan? There is the question of theimpact on the system as a whole. Some of thesechoices involve legal changes, such as how discrimi-nation rules would be remodeled, since we have todecide how much mandated savings would affectthe determination of whether plans are discrimi-nating any longer against low-income employees.Similarly, there are questions of portability, partici-pation, and employer design of benefit and paysystems.

■ The Question of RisksThere are a variety of risks specific to the plansthat have been advanced so far. One risk with themaintain benefits plan being considered by theSocial Security Advisory Council is the market riskthat comes from the government starting to investthis money in the stock market. Questions includewhat happens to bond interest rates, what happensto the bond market, and what happens if othergovernments start following this same logic andstart investing in the stock market?

If one thinks that one can arbitrage thefederal system, then why not have the federalgovernment borrow a trillion dollars and invest inforeign markets? Maybe we will not have anybudgetary problems at all, if we think we cansimply arbitrage to deal with them.

In terms of the privatization, the strictestprivatization plan considered by the Councilinvolves increases in government debt. My under-standing is that the nominal debt increases byabout $2 trillion at its maximum, or $650 billion inpresent value terms. This approach would haveeffects on the markets that need to be taken intoaccount.

There is a government risk in this plan as


76

well. There is a $410 basic benefit. That isabout two-thirds of the poverty level, as I under-stand it. If the money in these private accountsdoes not succeed in generating much in the way ofreturn, or individuals invest poorly, there is a highrisk that government will have to supplement thisbasic benefit.

■ ConclusionI think the type of modeling that EBRI is doing willbe useful in a number of ways. In some cases, it willidentify risk better than we have been able to dobefore. In the identification of risk it is not alwaysimportant that we get our estimates right; whetherthey are right or not, the models often tell us howto redesign reforms in ways to modify these risks.

For instance, if our risks are very sensitive to

life-expectancy—and I think they are—it is not sohard to think of a reform design that indexes theage of retirement for life expectancy. What thismodel tells us, regardless of whether the numbersare precise or not, is that a risk is present. If itappears to be large, we can think about remodelingSocial Security to deal with it.

Another example concerns individual invest-ment. If some people invest poorly, with little toshow from their private accounts, and also havebasic benefits that are well below poverty, govern-ment would have to back up a reformed SocialSecurity with a large welfare program. The model,therefore, warns us that we need to start examin-ing ways of regulating returns in a privatizedsystem. Chile is often given as an example of asuccessful privatization system, but Chile greatlyregulates minimum returns on privatized accounts.

77

Chapter 3

10■ IntroductionFirst, I want to salute the Employee BenefitResearch Institute (EBRI) and Dallas Salisbury fortheir splendid commitment to developing a newmodel that can evaluate the effects of variousapproaches to Social Security reform. It is aserious effort, responsive to the needs of both thepolicy community and the political decision makersthey advise, and it will be of inestimable value fordeveloping Social Security policy and fosteringinformed political debate about it.

I would like to pass over the technical issuesconcerning the model. These issues are of seriousinterest to econometricians and other academicspecialists, but Social Security reform is not anacademic exercise. It is a political enterprise thatwill touch virtually every American. Consequently,I would like to raise a few issues concerning, if youwill, the model’s conceptual base line; that is, issuesthat will help form the political predicate for thedebate over Social Security reform.

■ The EconomyThe first issue concerns our efforts to project theeffects of various reforms on the economy. Struc-tural reform of Social Security raises economicquestions about not only the impact of the fundingshortfalls expected under the current arrangementsbut also these arrangements’ other effects oncapital formation and other basic elements of theeconomy. The model should try to take account ofthese dynamics. Let me restate this point moreconcretely. The model should enable us to evaluatethe effects on the economy of the current system aswell as of various reforms. In this way, we can morefully understand the likely impact of variousreforms by evaluating them relative to not only thecurrent system but also to the economy, apart fromthe current system.

Issues Concerning the Modelby Robert Shapiro

■ Trust Fund AssetsThe second issue I would like to raise concerns howthe model will treat Social Security trust fundassets in evaluating the implications of both thecurrent system with its expected financing shortfalland various approaches to reform.

One of the strengths of the EBRI project is itscommitment to address Social Security reform notonly from the perspective of the system itself butalso in the larger context of overall governmentspending and revenue collections and from thelarger vantage of the Social Security system’s placein the economy. These are the very considerationsthat have led many reform advocates to proposesome element of prefunding or higher currentsaving in addressing the system’s coming fundingshortfall.

Prefunding part of future liabilities wasalso an essential element of the 1983 Social Secu-rity reforms. Yet, I believe most of us could agreethat the prefunding that has occurred under thosereforms is of a very peculiar sort, from the largerperspectives of the entire government and theeconomy. From these vantages, the resourcesrepresented by the trust fund balances have notbeen saved in the ways that are conventionallyrecognized by economics or accountancy. Thesecurities held by the trust fund represent aneconomic asset for the Social Security Administra-tion but a corresponding economic liability for therest of government—and one that will entailadditional revenues, spending reductions, orborrowing of precisely the same kind and extent aswould be required if there were no trust fund. Thetrust fund’s assets also represent a correspondingliability for the economy, because the surplusrevenues they represent have been used not toincrease the capital stock but only to financegovernment activities.


78

We cannot sensibly evaluate the economicimplications of reform if we do not acknowledgethat financing based on these trust fund assets, ascurrently planned for the years 2013 to 2030, willinvolve additional borrowing, spending reductions,or tax increases equal to the face value of theseassets.

Yet, in most current discussions, and in theAdvisory Council’s three reform proposals, theeconomic value of these trust fund assets is simplyassumed. I urge EBRI to address the trust tundissue independently and to incorporate the eco-nomic implications in its model.

■ Modeling Future PerformanceThird, I would like to raise the issue of our abilityto sensibly model the future performance of finan-cial asset and debt markets over substantialperiods, as would seem to be required in order toevaluate the implications of various proposals forpartial privatization or government investment inprivate asset markets. It is the nature of thesemarkets that they capitalize countless unantici-pated economic events and developments, so that inorder to project their future performance we wouldhave to be able to anticipate the equilibrium effectsof such events and developments. Yet, if they couldbe anticipated, it would disable these markets.

At a concrete level, try to imagine, if you can,what a model of our equity and debt marketsconstructed in the 1960s would have projected forthe 1970s, 1980s, and 1990s, relative to the actualperformance of these markets.

More particularly, I am concerned about theapplicability of whatever conclusions are arrived at,concerning future average market rates of return,to periods of a few decades and to different classesof investors. These markets have demonstratedenormous variation in rates of return from year toyear, and these variations if continued would haveenormous effects on the resources available todifferent cohorts of retirees under any form ofprivatization. Robert Shiller at Yale, for example,has documented 15 years in the last 70, includingseven years in the 1970s and 1980s, in which the

inflation-adjusted value of the Standard & Poors500 Index was at least 40 percent lower than it hadbeen 10 years previously.

These markets also display enormous varia-tions in rates of return among different investorsand classes of investors. There have been twostudies of these variations, and both suggest thatinvestors’ rates of return are generally associatedwith income. In particular, high-income investorsreport rates of return roughly three times as greatas those reported by moderate-income investors, sothat high-income people earn higher than averagerates of return and moderate-income people earnlower than average rates of return. I seriouslyquestion, therefore, the usefulness of any assump-tion about a future average market rate of returnfor evaluating the social implications ofprivatization reform proposals.

At a minimum, the model should take accountof the problematic character of our ability to projectwith any confidence the impact of private invest-ment over periods of time and among differentgroups of investors.

■ ConclusionFinally, I want to suggest an additional dimensionfor the EBRI project, expanding its conceptual baseline. One of the project’s intriguing elements is itscommitment to provide a means of evaluating thenormative implications of various reform proposals,with respect to both the distribution of benefits andtaxes, and the political character of the newarrangements they would entail. Given this ambi-tion, the project could benefit from a systematicevaluation of public opinion, probing people’sresponses to various alternatives and their implica-tions. By documenting the schedule of values whichthe public holds in these areas, the project couldenrich the debate about the normative implicationsof the various reforms.

In closing, once again I want to salute EBRIand its project. I hope that by raising certainconceptual issues, I have provided a small contribu-tion to this splendid and important effort.

79

Chapter 3

11Social Security Reform: Gaps inPerceptionby Susan Dentzer

■ IntroductionThe Advisory Commission to Study the ConsumerPrice Index (CPI), headed by Stanford economistMichael Boskin says that the overstatementimplicit in the CPI is about 1.1 percentage points ayear. The commission further suggests that fullymore than one-half of this overstatement is due tothe fact that the CPI is incapable of adjustingadequately for changes in the quality of goods, andin particular the quality of new products. Anestimated .6 percentage points of that 1.1 percent-age points of overstatement is attributed to thisphenomenon.

However, consider how the commissionreached this conclusion. For example, it sampled27 categories of goods in the CPI. It concluded thatthere was not much of a bias to speak of in 7 of the27 categories, and there was substantial upwardbias in the other 20. The commission concludedthat apparel has actually been improving markedlyin quality and that, as of about 1985 onward, theoverstatement of the inflation in the apparel sectorof the CPI has been one percentage point per year.

As a person who has occasionally done someshopping, the overwhelming impression I do nothave when I shop is that the quality of apparel isimproving by 1 percent a year relative to the price.

You get a sense, as one of the commissionmembers said, that the precision of these numbersis not to be taken to the third decimal point. It isgoing to be a very long process as the Bureau ofLabor Statistics systematically examines thesekinds of suppositions. The commission’s perspectivethat, in every day and every way, things are gettingbetter, is probably not to be taken at face value.

I think there is a key object lesson for us here,as we contemplate the future prospects for SocialSecurity reform. As the commission’s good workdemonstrates, there is a large, potential perceptiongap between the work of individuals who are

familiar with these issues’ abstruse aspects and theway the issues are likely to be perceived by thepublic.

A recent preliminary analysis of the BoskinCommission Report that appeared in the Washing-ton Post pointed out that telling Social Securitybeneficiaries that benefits for an average couplewould be cut next year by $200 or so as a conse-quence of the overstatement of the CPI will not beeasy. Merely saying that it’s quite all right, becauseProfessor Boskin says so, is probably not going tofly with the general public.

One cannot spend as much time going toSocial Security privatization conferences as I doand not conclude that this perception gap is hugeand is likely to become even larger before it beginsto close in the years ahead. For example, those ofus who lived through health care reform went tomany a conference where we discussed issues suchas the impact of regional health care alliances andrisk adjustment among plans and listened to wisepeople model the potential risk adjustment mecha-nisms that could be undertaken to even out thedifferences among plans. Nobody sat down andmodeled for us the impact of “Harry and Louise”commercials, which, in the final analysis, proved tobe much more instrumental in derailing health carereform than anybody’s lingering concerns abouthow we would do risk adjustment among the plans.

Let me introduce what I think are some of thekey reality gaps and perception gaps that I per-ceive, particularly in my work as a reporter cover-ing these issues and also as an observer of the waythese things have played out, as in the case ofhealth care reform, over time.

■ Social Security Participants asInvestors

With respect to partial privatization of the Social


80

Security system, the first issue is the education oftomorrow’s investors who are going to make thesewise investment decisions that may confront themin the future. Based on the tenor of much of thereader mail I receive—particularly when I write acolumn on Social Security—I have my doubts that alarge segment of the population would be able tohandle the investment choices involved. I amfurther buttressed in this view by statistics show-ing that 70 percent of Americans now get theirprimary news from television.

I don’t expect Oprah, let alone, perhaps, eventhe CBS Evening News, to go into graphic detailwhen it comes time to explain annual investmentchoices that individuals may need to make or notmake in order to maximize their investmentreturns.

Larry Summers, the Deputy Treasury Secre-tary, is fond of tweaking the financial press forrunning volumes of articles on how to pick amutual fund with the best return without evernoting that, in fact, almost no one beats the marketover time. If you were really smart and wanted tosave yourself a lot of headache, you could probablyjust put your money in a stock market index fundand not worry about the emerging markets funds’return relative to some other funds’ return, etcetera.

My experience suggests that many individu-als are not going to be any match for the ingeniousmarketing that Wall Street and other investmenthouses are capable of, and I frankly have greatdoubts about the ability of many of the peoplewhom I know to be among my readers to chooseamong investments and not get corralled intodecisions that could be a large error for them overtime.

A lot of people blithely point to statisticsabout how smart people are becoming as theyinvest in 401(k) plans. Again, I think there is a bigreality gap that we have yet to fully contemplate.

■ Annuities MarketsAnother perception gap or reality gap came uprecently at a Council on Foreign Relations confer-ence on privatization of Social Security and otherpension funds around the world. There was discus-sion about how the linchpin of a partialprivatization effort would be the availability of

some kind of well-developed annuities market sothat people could cash out their personal securityaccounts at age of 65, or whatever, convert them toan annuity, and then have that money flow to themover time. Can we be confident that an appropriateannuities market would, in fact, materialize?

Those familiar with this issue know thatthere is a much debate among economists aboutwhether this is possible. There are certain asymme-tries of information between individuals and sellersin the current annuities market, and it is not clearthat the pricing of annuities will operate in favor ofthe buyer.

When this issue was brought up at theCouncil on Foreign Relations conference, thefinancial market participants present simplystated, “Well, build it and it will come.” That is, ifyou partially privatize, a flourishing annuitiesmarket will develop, and we can conquer all marketopportunities. We have demonstrated our ability todo that. And everything will work out fine.

This may happen. However, important issuesare involved in selling people a plan that is predi-cated on having reliable returns over the course ofthe investment period and also having a reliableway of securing the investment that has accrued upto the date of retirement. Merely hearing fromfinancial market participants that these issues willtake care of themselves does not increase my levelof confidence that this is a salable plan.

■ Raising TaxesWith regard to selling these reform proposals to thepublic, there is yet another perception gap and areality gap. It pertains to the partial privatizationproposal put forward in the Advisory Councilreport. If, in fact, we deem the long-term shortfallof Social Security to be roughly 2.17 percent ofpayroll—or 4 percent if the life expectancy statis-tics are really as bad as we hear—then the personalsavings account (PSA) plan requires at least a one-and-a-half percent payroll tax hike to retire the so-called liberty bonds that will be used to finance thetransition. Is a 1.5 percent payroll tax hike to retireLiberty bonds inherently more salable than a2.17 percent straight payroll tax increase?

This does not sound inherently more salableto me. If members of generation X are going to haveto pay twice for our retirement and their own, do

81

Chapter 3

they particularly care how they pay twice for it? Itis not clear to me that one particular proposal isgoing to be more salable than another.

■ The Retirement AgeI would like to make two further points on thereality and perception gaps. The whole question ofextending the retirement age is one that those of usin the press and in policy circles discuss in waysthat simply do not take into account how this islikely to play out in people’s lives.

Again, at the recent Council on ForeignRelations conference, there was a unanimoussentiment that the normal retirement age has to beraised. Why not age 70? After this conference wasover, I wanted to take a survey asking how manypeople in that room actually will be compelled, forfinancial reasons, even to work beyond age 60.Given the tier of people who tend to go to meetingsof this type, I would venture to say that nobodywould have fallen into that category.

In contrast, to take an example close to home,my son’s babysitter, now age 65, whose SocialSecurity benefit is $500 a month, will probably notonly have to take care of my son until he is incollege but presumably also care for my grandchil-dren in order to supplement her Social Securitypayment. Her expectations of retirement at any ageare probably nonexistent.

Even if such people understand that they aregoing to have to work for a long time beyond whatwe would consider normal retirement age, tellingthem up front that their retirement age is auto-matically going to be raised to age 70 will be a verydifficult political argument indeed.

Consider this together with James Smith’scomments to the effect that the system is alreadygoing to be less progressive over time for those inthe lower socio-economic strata.1 It would suggestthat the consequences of perceived social injusticefor different classes of individuals are far greaterthan we currently perceive. In fact, this whole issueof class differences, socioeconomic differences, and

the way different categories of the population willperceive reform proposals has also not receivedsufficient attention and is probably impossible tofactor into any kind of model.

■ Medicare and Social SecurityThe final perception gap I perceive is over thissimple proposition: why is everyone paying so muchattention to Social Security reform right now, whenMedicare reform is obviously the bigger fish thatneeds to be fried? I have a feeling that if there wereas many people—such as those on Wall Street—who had as great an investment in Medicare reformas they do in Social Security reform, I would spenda lot more of my time going to Medicare reformconferences these days than I do going to SocialSecurity privatization conferences.

Gene Steuerle2 and others have calculatedthat the Medicare long-term financial shortfall, ifthe current projections hold, is roughly three timesas great as the Social Security long-term shortfall.That suggests to me that it is three times moreimportant to start talking about Medicare reformthan it is to talk about Social Security.

Even if we don’t do that, we should be discuss-ing the two programs in tandem. I cannot under-stand why we ever discuss them as separatepropositions, since they are so clearly linked at somany levels. They are linked on the revenue side,with respect to financing by the payroll tax. Theyare linked, quite obviously on the expenditure side,with respect to not only federal expenditures butalso to individual household expenditures.

It is a curiosity to me that we don’t spendmore time considering these programs together. Iwas heartened by Dallas Salisbury’s comments thatthe effects of Medicare changes and corporatechanges in retiree medical plans will conceivably beadded into this modeling process in the future.Unless they are, I would suggest that we are notengaged in a particularly meaningful effort inevaluating Social Security reform on its own.

1 See James Smith, “A Comment on Mortality As-sumptions, Low Wealth, and the Need for MoreSaving,” in this volume.

2 See Gene Steuerle, “What Can We Learn fromModeling Social Security?” in this volume.


82

83

Chapter 3

Investment Implications of SocialSecurity Reform: More Questionsthan Answers

12by Girard Miller

■ IntroductionThe nonprofit organization that I head handlesretirement plan administration issues for state andlocal governments. We are predominantly a definedcontribution plan provider. Because of that, whenwe first looked at Social Security reform, we said,“This is great.” We are like everybody else in thefinancial services sector, trying to tap some newgold mine that will provide an unlimited number ofassets to fuel the demand for plan administrationand investment management.

As I became more deeply involved in SocialSecurity reform, however, I found that it is muchmore complex than most of us would have thoughta short time ago. I think we all need to address thisissue at a deeper level. I am going to discussseveral nontrivial questions and challenge you on acouple of issues as well.

When considering the fate of Social Security, Iam reminded of the story of the English farmer whohad the only horse in the county. One day thefamily came in yelling, “Something terrible hashappened. The horse has run away. Woe to all ofus.” They ran around town with scowls on theirfaces all day. The farmer, however said, “Don’t bealarmed. It’s too soon to tell.”

Sure enough, that evening, the horse cameback with three wild stallions. Of course, the familywas jubilant. “We’re now the wealthiest people inall of this side of England.” And the farmer againsaid, “Too soon to tell.” The next day, his son tookout one of the wild stallions, ran it down to the endof the farm yard, and ran over a fence. The horsewent head over heels, the kid broke his leg in threedifferent places, ended up in the hospital in trac-tion, and, of course, this was terrible. Those whowere so happy in the morning now were morosebecause the harvest was due to come in, the soncouldn’t help, and life was terrible. Again, the

farmer said, “It’s too soon to tell.” The next day, thearmy came into the land, conscripted every 18-year-old who was able to serve in the army, took themacross to the other side of the world, where they allwere killed—except for the farmer’s son, who washospitalized that day.

As we move into the world of Social Securityreform, privatization and the rest, we should bemindful of the law of unintended consequences andof the fact the future is much more uncertain thanthe models suggest.

■ Workers as InvestorsWith that admonition about the perils of prematurejudgments, I’m going to start with one of thenontrivial issues. If 200 million or more people inthe United States are permitted to invest their ownretirement money, could they do it successfully? Ithink this is a fundamental question and one onwhich we will eventually need to spend a lot oftime.

I can share with you information based onexperience. Our organization, like many in oursector, had for some time been dominated by whatwe call stable-value investments—insurancecontracts, and the like. Ours is a voluntary plan formost people; however, it represents the average citymanager’s entire life savings. We have 260,000people in this program; chart 12.1 shows the mix ofa different series of voluntary retirement savingsamong state and local government workers.

In 1990, 70 percent or more of the money wasin stable-value, or GIC-type, contracts. Thisnumber has changed dramatically. Some of thechange reflects the great bull market of the 1990s,but a lot of it has to do with our concerted effort inemployee education. We have developed the tools tohelp the average worker in state and local govern-ment better understand how to invest money


84

1990 1991 1992 1993 1994 1995 1996

Fixed

Equity

Plus

100

80

60

40

20

0

prudently. We have developed model portfoliopackages. I think it can be concluded—afterconsidering the numbers—that our program nowlooks much more like a traditional defined benefitpension plan in terms of the overall asset mix. Yetthe plan even includes a large number of retirees.

So there is some hope that employees can, infact, be trained, and I think the work that DallasSalisbury and the others have been doing on theAmerican Savings Education Council is indicativeof that. As chart 12.2 shows, the allocation ofrecently contributed money now resembles thesame mix that you would expect a private invest-ment advisor to provide to wealthy individuals.

These are average numbers across a wholesystem. One thing I found interesting about theEmployee Benefit Research Institute (EBRI) modelis that it basically says, “100 percent is going to gointo equities in the early stage of life, and23 percent going into equities during retirement.”This actually is not that far from our particulargroup’s actual experience these days on the basis ofnew money in.

Next, I return to my original question: If200 million people are turned loose to invest theirown money, could they effectively manage it? Evenif the answer is “Yes,” I think we then have to ask,“Will individual ownership of these assets necessar-ily be the best policy?” People will split on this

issue, based on economic philosophy, politicalphilosophy, or whether or not they are involved inthe financial management industry.

Economic science does not tell us much aboutthis. There are major long-term implications ofthrift that we need to think through as a country,such as whether or not there will be behavioralshifts. I don’t think the EBRI model is going to beable to predict the implications of telling massivenumbers of American workers that the nest egg intheir personal savings account will be their numberone source of future retirement security. We need tothink about this.

■ CostsAnother question to ask is, will not the SocialSecurity safety net itself—whatever is left of it ifprivatization occurs—need to be extended to covereither losses on ill-advised investments or retireecohort-specific market losses? I am referring topeople who either bungle it in terms of theirindividual investment behavior, or the chance thatthose at the end of the baby boom cohort mayexperience a repetition of the recent Japanese stockmarket decline and its impact on retirees. Both ofthese scenarios could throw many people into asocial safety net. We don’t have much in thedynamics of the EBRI model at this time that

Chart 12.1Asset Mix, 1990–1996

Perc

enta

ge

Source: Girard Miller.

85

Chapter 3

Chart 12.2Participants’ Contribution, 1990–1996

100

80

60

40

20

0

addresses this fundamental question. I think it isone we need to understand better.

■ AdministrationThe next question goes to the administrators ofSocial Security. Will there be a national defaultinvestment product? That is an administrativeissue about which we have heard very little intelli-gent discussion.

Concerning administration, at a practicallevel, what would be the expense ratio of theinvestment products, and what would be the planadministration fee, for offering all of this? We havegood statistical data from the Society of PlanAdministrators and Record Keepers on the cost of401(k) plans, 401(a) plans, and 457 plans. We cangive you these numbers, and they would probablydrop by 50 percent in the five-year period that Ipredict it would take before we could implementprivatization measures—even if Congress could bepersuaded to start tomorrow.

Aside from that issue, we do not have num-bers concerning what it would take in terms ofrelative costs to administer itinerant workers, part-time people, and others. My organization’s planprobably has more small deposits and smallaccounts than any organization in the country. Weknow that, as a result, our average costs are higher

than those of the mainstream mutual fund andrecord-keeping providers. When we universalize tothat segment of the population that lacks substan-tial wealth and assets, the average administrativecosts for plan administration fundamentally will behigher than the costs the industry experiences now.Building in those expenses could raise a fundamen-tal policy issue.

Even more importantly, we need to addressthe issue of the cost of the investment vehicle. Arewe talking about three basis points for a Barclay’sindex fund, or are we talking about 90 basis pointsfor Fidelity Magellan?

■ RegulationFrom the standpoint of fundamental social policy,this leads to my next nontrivial question: “Who isgoing to regulate, and how will we regulate andcontrol the investment products under aprivatization scheme?” I think those at the Na-tional Taxpayers Union would say that theyenvision some kind of “guard rails” on the system.This implies that there would be a knowledgeablefederal body that is able to determine what invest-ment products are appropriate for a privatizedscheme. And that, again, is not trivial.

1990 1991 1992 1993 1994 1995 1996%

%

%

%

%

%

Fixed

Equity

PlusPerc

enta

ge

as of 6/30/96Source: Girard Miller.


86

■ Macroeconomic ConsiderationsFinally, I want to address a macroeconomic ques-tion: the issue of whether there truly will be ameasurable effect on the returns and the risks ofequity capital if we popularize the ownership ofindividual equity accounts through a fundamentalchange in Social Security. That is not a trivialquestion.

Such an initiative would change the supply ofcapital going into the equity markets. A practicalmacroeconomist could say that the essential long-term effect of that could be an eventual reduction ofreturn on capital; that could have, I suspect, asignificant influence on what measured return youexpect from equity investments over the long term.This is a very interesting challenge, and moreimportantly, one that affects the stochastical andMonte Carlo models because it will change therisks.

Consider Japan, one of the superior societiesin the world when it comes to savings, and reflectfor a moment on the risk and return characteristicsof the Japanese stock market in the last 15 years.That is, again, not a trivial question.

■ ConclusionI have introduced more problems than answers, butthey are fundamental ones. I want to congratulateEBRI for at least opening the dialogue. I concludewith one more challenging thought: Some peoplesay that, without change, the Social Securitysystem will be bankrupt by 2030. However, a largepart of that deficit can be changed without any-thing related to the privatization of Social Security

accounts—with simple reform.For example, in talking to people around the

country, when I mention raising the eligible retire-ment age from 65 to 70, I see nothing but headsnodding in agreement. I don’t think this conceptwould be as difficult to implement as some of thepoliticos and pollsters suggest. However, a higherretirement age of the Old Age, Survivors, andDisability Insurance (OASDI) program would havea profound impact on employers and also on thosewho administer retirement plans. It would changethe behavior of individuals ages 62 to 70; forexample, they might end up spending down moneyin other private accounts to fund early retirement.

Finally, we need to engage in a dialogue,similar to this one, with respect to Medicare—asystem that will be bankrupt much earlier thanSocial Security. Some of us would say that, if thepolitical priorities were straight, we would beaddressing that issue first and Social Securitysecond.

Certainly, there are many issues that relate toretiree health care. We should consider whatpotentially we should be doing with sec. 401(h)accounts in the current code. These vehicles couldprovide a means for employers and employees tosave money to supplement, or complement, theexisting Medicare system and to provide, for thefirst time, a market-based control on the demandfor retiree health care. That is completely missingin this country today. So I hope you now canappreciate that we still have more questions thansolutions. But, fortunately, we are on the righttrack in studying them now, well in advance of anactual political move. Just remember that it is toosoon to tell.

87

Chapter 3

13■ IntroductionI am in the rather unenviable position of being astatus quo-er. It is always so much easier to haveradical thoughts, to show that you are reallythinking, than merely to say, “Keep on doing whatyou’re doing. It’s fine.” But to paraphrase a commonexpression, I don’t think Social Security is broke.Don’t throw it away. I think that it may need somechange, and this would be prudent to do and wouldhelp restore public confidence. However, this shouldnot be done radically or quickly but rather in adeferred and very gradual manner. None of usknows what the future will bring, and if action istaken in a deferred, gradual manner, it is mucheasier to change course later than if you act now inhaste.

Robert Friedland1 described excellently how,at the present time, many people are completelydependent on Social Security, and many are largelydependent on it, while very few are partiallydependent on their children, and virtually none arewholly dependent on their children. In contrast, inthe 1930s, those latter proportions were muchhigher. Social Security is doing an excellent job, solet us not throw it away.

Today, people talk a lot about how greatinvestment accounts are, and I would not deny thatthey are. However, I don’t think that they take theplace of Social Security, or even part of it. Also, theword “privatization” sounds good to people, like“motherhood” and “apple pie.” But if they reallythought about what it means, they might not be soenthusiastic.

And I cannot believe that most people—asidefrom those who are well-educated and financially

Social Security: Myths and Realitiesby Robert J. Myers

astute—want to manage their money. What theyare interested in is having at least a basic floor ofeconomic protection, which is what Social Securityoffers, and then building on it as they choose. Andmaybe they don’t even choose to do it at all, but atleast, then, they have that freedom.

Relying wholly on individual savings ac-counts, with individuals making their own invest-ment decisions, sounds good to the financial elite.Choice is fine, but in a highly industrializedcountry like ours, people do not have completechoice. We do not have a choice whether or not tostop at a red traffic light. That choice has beenmade for us.

Furthermore, many people are saying, “Let’schange Social Security. We can do all sorts of thingsfor the economy through Social Security.” I don’tthink that is the case at all. I do not think SocialSecurity’s purpose is to solve all national problems.

Some years ago, Social Security was criticizedbecause it did not do enough to provide equalemployment opportunities for women, or equalemployment opportunities for minorities. That wasnot its role. I think that Social Security did developmore equal treatment, but these other elements areoutside of Social Security. We should not expect toomuch of the program.

Today, people are saying, “ We must changeSocial Security. We need more economic growth. Weneed more savings.” I won’t argue about that. Thatis not my field. Undoubtedly, we do. But I do notthink that is Social Security’s responsibility.

■ Myths about the SystemThere are several myths about the original SocialSecurity program of 1935 that still prevail and tosome extent affect current thinking. One is that theoriginal system was fully funded, similar to anygood private pension plan. Or, at the oppositeextreme, some say that it has always operated on a

1 See David V. Bryce and Robert B. Friedland, “AnOverview of the Origins of Social Security,” in thisvolume.


88

pay-as-you go basis. Obviously, it cannot be both. Infact, Social Security was created as a partiallyfunded system intended ultimately to provideinvestment income that would finance about25 percent of the costs of the program, by buildingand maintaining a large fund.

The second myth about the original programwas that it operated completely on an individual-equity basis. “Everybody would receive just whatthey paid for. This was a real insurance system,just like an insurance company would sell.” Again,this was not true. The original benefits wereweighted heavily toward the lower paid and towardthe people who would retire after contributing foronly a few years. True, there was a provisionstating that everyone was guaranteed to get backthe employee taxes, plus a small allowance forinterest, but that is not true individual equity. Thatis just a move toward individual equity, and theprovision was eliminated in the 1939 Amendments.

Still another widespread fallacy is that theoriginal act was intended to supplement individualefforts and supplement private pension plans. Thefact is there weren’t a great many private pensionplans in the early 1930s, nor were there a lot ofsavings, because they had been wiped out in theGreat Depression. Instead, Social Security was putin as a basic guarantee, a basic floor of protection.

It has also been said that Social Security wasestablished to reduce unemployment by encourag-ing people to retire and make jobs available foryounger people. While it may have had a veryminor effect on employment patterns, if that wasthe intention, it was a poor plan, because themonthly benefits were not scheduled go into effectuntil seven years after it was enacted and thenthey amounted to only about 15 percent of earn-ings. How could that persuade people to retirevoluntarily?

Several other myths are currently enteringthe debate about radically changing Social Security.One is that the program absolutely will go bank-rupt in the year 2029, and that’s it. It is alleged tobe an inflexible program that cannot be changed.

As has been frequently discussed, there areseveral actuarial cost estimates for the SocialSecurity program. According to the low-cost esti-mate, which contains reasonable assumptions, theprogram is in great shape forever—not only for75 years, but beyond that. I don’t think this is too

likely, but it is possible. Nonetheless, I think thatwe should soon take some actions that I willdescribe later. But it isn’t that we know that wemust take action, but rather that it is desirable todo so.

Also frequently discussed is what I call theactuarial concept, “Do people get their money’sworth?” Or what about their rates of return? Manypeople believe these considerations to be veryimportant.

These considerations are interesting, but theyare not relevant. That is not the basic principle ofthe system, which is to provide a floor of economicprotection or income maintenance. Whether peopleget their money’s worth or not is not material, anymore than it is material as far as school taxes areconcerned.

Another widely discussed matter is that thetrust fund investments are worthless IOUs, be-cause the money has been spent. However, the factis that if the trust funds had not bought thesebonds, somebody else would have had to buy them,and those bonds would have had the same validityand would likewise be part of the national debt.The fact that the bonds are not marketable isirrelevant, too, because much of the public debtheld by private individuals is not marketable—e.g.,the Series E bonds. They are redeemable at ascheduled value at any time, but they cannot besold in the open market.

Furthermore, it has been said that theinterest on the trust-fund investments will not beused for many years. This is not true. Just a fewdays ago, when some 43 million checks were mailedout or were credited to people’s bank accounts,securities were redeemed by the trust funds—despite the popular belief that securities will not beredeemed until 2019 and after, and this will thenbe a great problem. Securities are redeemed everymonth. Moreover, part of the benefits that go outeach month come from interest—namely, theaccrued interest on the securities that are re-deemed. Looking at it from the standpoint of aneconomist, with money being fungible, you can’t saywhere the benefits came from—from money derivedfrom payroll taxes or from interest on the investedassets. And, incidentally, in a sense, although itdoesn’t mean anything very significant, the intereston the investments is now paid by check, not by abook entry or bookkeeping.

89

Chapter 3

■ ModelingI believe that, in connection with modeling, we aretrying to be too scientific. I don’t believe in stochas-tic modeling one bit. It is a great intellectualexercise, interesting to the technically minded. But,as a practical matter, I don’t think whether peopleget their money’s worth is relevant, and certainlynot any more relevant than is the case in connec-tion with school taxes. I certainly agree withMarilyn Moon’s statement that, because the dataare so variable, it is difficult to give much credenceto the results that come out of this little black box.2

■ PrivatizationAs to privatization of the trust funds’ assets, I thinkthat this would be a bad idea. I think it desirablethat the trust funds not be too large. I am anadvocate of pay-as-you-go financing. Building uplarge trust fund assets is dangerous politically andeconomically, and the government should stay outof ownership of private industry.

Concerning the privatization of the benefitstructure, as two of the three Advisory Councilgroups have recommended, it is interesting that itsadvocates approach it with various differentagendas. Some people sincerely believe thatprivatization will provide better benefits foreverybody—or, in other words, perpetual motion.Others view it from another angle: “Wouldn’t thisbe good for my business? Look at all the additionalinvestment business we will have.” Still otherpeople support privatization because, for years,they have been philosophically strictly opposed toSocial Security, or to the government doing any-thing for anybody. In other words, they believe“Everybody should fend for themselves and, if theydon’t, that’s their fault. Let them starve on thestreets, but my money is mine, and if they don’thave money, too bad.”

One of the many problems with privatizationis annuitization. Should it be voluntary or compul-sory? Either way, there are problems. If it is doneon a voluntary basis, it works out unfairly forwomen, because they have longer life expectancy. Italso works out unfairly for one-worker families—

and there are still many of these families in thecountry—because they have to split the accumula-tion between two people; it does not go to a short-lived male.

Also, there is the problem of people withoutdependents who build up a large account and arenot in good health, or perhaps die just beforeretirement age. What economic or social purposehas that large accumulation served?

Still another problem is what to do about thesmall accounts. Privatization works out well forsteady workers with average or high earnings whomake large contributions. Mutual funds andindividual accounts cannot efficiently handle smalldriblets of money coming in intermittently andthen keeping track of the investment performanceand reporting on it every three months. Therewould be very high administrative expenses.

People sometimes gloss over the difficultieswith privatization involving the increased tax rates(which may not be popular), increased nationaldebt, and increased budget deficits. I don’t see howthis approach can possibly get anywhere in aCongress that claims it is bound to try to balancethe budget in seven years, not make it larger.

The proposals that I have seen do not give anyindication that they have smoothly merged disabil-ity benefits and the young-survivor benefits withthe retirement benefits. I think there could besharp notches and glitches among people who drawdown their accumulation at the wrong moment ordie just before retirement age or just afterwards.On the other hand, the present Social Securityprogram has, generally speaking, very smoothjunctions.

■ Maintaining the SystemI will end with my proposal. What I think ought tobe done is to maintain the Social Security systemas we have known it, as it has operated for60 years, but make changes in it that are reason-able, affordable, and logical.

One change is to raise the normal retirementage. That reduces the benefit cost, but I don’tconsider it a “real” benefit reduction in view of thevalue of the lifetime benefits. If the age were raisedslowly to 70, as it is now, starting in 2003, butgetting to 70 by 2037, people age 70 in 2037 wouldprobably live about as long as people retiring at age2 See Marilyn Moon, “Social Security Reform: Beyond

the Models,” in this volume.


90

65 today. Thus, that is not a “real” benefit cut.It reduces benefit costs on a year-by-year basis, butnot on a lifetime basis.

The other part of this proposal would be thetraditional one of not only reducing benefit costsbut also raising income. I would increase the taxrate, beginning in 2015, by 0.3 percent each on theemployer and the employee, and then impose asimilar increase in 2020, 2025, and 2030, a totalincrease of 1.2 percent on each. With any sort ofreal growth in this country, people’s real incomeswill continue to rise in those years. Further, even ifsome increased portion of earnings is taken awayfor higher medical costs, I think that real incomeswill still rise.

And even if real incomes do not rise, doAmericans have to be so unaltruistic? Do they haveto be so concerned with constantly rising incomes,and have five cars in every garage, and threetelevision sets in every room? We have a very highstandard of living in this country for most people.Something ought to be done with regard to thosewhose living standard is inadequate. But themajority ought to be satisfied with what they have,or at least with smaller increases than would occurif they did not consider the needs of others.

I also think that it might be desirable toestablish mandatory individual accounts that

would be completely separate from Social Security,and see how it works. And again, leave out the verylowest paid, because it is not efficient to includethem, and it would do little for them.

Finally, I would reallocate the Social Securitytax rates beginning next year and cut the SocialSecurity rate by 0.6 percent each on the employerand the employee, and put that money into theMedicare Hospital Insurance (HI) Trust Fund. Bydoing this, the Social Security trust-fund reservewould remain at a satisfactory contingency-reservebasis. This would give Medicare HI an extendedlifetime during which a solution to its long-rangefinancing problems could be worked out. I thinkthat the eventual solution for the HI problem willinvolve higher costs, and people can bear it. Atleast, my proposal would remove the problem fromthe immediate front burner.

■ ConclusionWhat I propose would solve several problems, andit would maintain what has been, and still is, ahighly successful program. It would not put us indanger of a new, radical solution that might createa terrible mess and create all sorts of problems thatwe cannot imagine at the moment.

91

14The Maintain Benefits Plan forSocial Securityby Robert M. Ball

■ IntroductionMuch to my surprise, I found my participation inthe 1996 Social Security Advisory Council to bequite an educational experience. Having beenassociated with some five earlier councils as well ashaving worked for Social Security for 30 years, Ihad thought that my positions on Social Securitywere pretty well set. But out of the discussions inthis council, I actually came to adopt one importantproposal—one that was completely new for me. Iwould like to explain to you how I came to advocateinvesting some of the Social Security build-updirectly in passively managed private equitiesindexed to the broad market.

Mainly, the reason for my advocacy is that Ibecame convinced that we had to do somethingabout the perception of younger people who believethey are not getting a good deal under SocialSecurity. And at the same time, I was convincedthat one way of attacking this problem—individualprivate investment taking the place of all or part ofSocial Security—was a bad idea.

I doubt that most young people believe theyare not going to get any benefits from SocialSecurity; but certainly a very large number of themhave concluded that Social Security is not a gooddeal for their generation. The point is not whether Ibelieve “money’s worth” and the “rate of return”should be major criteria for judging Social Security.The point is that if Social Security is to survive,those who are going to vote in the future need tounderstand and support it, and they believe thatmoney’s worth—the ratio of benefits to contribu-tions—is an important criterion in decidingwhether or not the system is fair.

So I started to look at this issue more seri-ously than I had, particularly as groups within thecouncil started to argue that the way to improve“money’s worth” for younger people would be toprivatize, or partly privatize, Social Security—that

is, to take some of the money that goes into SocialSecurity and give it back to workers to invest inpersonal accounts. The argument was that presum-ably individuals would invest some of this money inprivate equities and, on average, this would resultin bigger returns than Social Security investments,which are limited to the returns payable on long-term government bonds.

■ The Problems With IndividualAccounts

It is true that generally this procedure might wellimprove the return on contributions, but as Istudied the proposals for individual accounts, I sawmany other problems with them. One is that thisapproach greatly reduces benefits in the govern-ment part of the program (a 30 percent reduction inthe average individual account (IA) plan and muchmore in the personal security account (PSA) plan).Thus, under these plans what you have left in thegovernment part is much worse than the currentplan in terms of money’s worth for the very peoplewhose status you are trying to improve.

And you cannot expect those covered by SocialSecurity to look always at the total program—whatthey get from their own individual accounts to-gether with Social Security. They will look at theresidual Social Security program separately. Overtime, those average and above-average earners whoturn out to be successful investors in their ownindividual accounts will see these accounts gettingbetter and the residual Social Security programgetting worse. Increasingly many may begin toargue: “Why can’t more of the contributions thatare going into Social Security go into my individualaccount? Why not just drop Social Security?”

That is why I believe that these two-partplans with individual savings accounts have inthem the seeds of dissolution for the whole Social


92

Security program. The redistribution of income thathas made Social Security such a successful antipov-erty program and such a successful program forlow-income people could disappear. Those who aredoing better than average in their individualaccounts may not be much concerned with thosewho are doing worse than average. But if they loseinterest in the traditional program, by defaultSocial Security could become a low-income pro-gram—and end up with the resulting weak politicalsupport common to all low-income programs.

And there are many other problems withpartial privatization plans. For one, I doubtwhether the savings would last until retirement.The idea behind individual accounts is to empha-size that they belong to the individual. As peopleface financial crises during their working careers,Congress probably would allow them access to thebuild-up in their accounts, even though in theory itis all supposed to be held for retirement. Butconsider an individual who is still unemployed afterhaving exhausted unemployment insurance; is theCongress going to say that person has to seek reliefregardless of the size of a personally held account?What about major medical expenses? Or the desireto have a starter house? Or education for one’schildren? If it is the individual’s own money, theCongress will be under great pressure to make itavailable.

Another problem is that both of theseprivatization plans in the Advisory Council Reportsubstantially increase the compulsory payments(taxes) that people have to make. The IA planincreases deductions from workers’ earnings by1.6 percentage points, which is then put into theindividual account. The goal is that the 30 percentaverage reductions in the residual Social Securitybenefits on average will be compensated withreturns from individual investment. That maywork—on average. But what about all those whosereturns will be less than average; or what aboutthose who may even lose all or part of the princi-pal? The basic tier of our traditional three-tier planfor retirement income should continue to be SocialSecurity, a defined benefit plan on which people canrely, without dependence on returns from indi-vidual investment. Then supplementary savingsand even supplementary pension plans, the secondand third tiers, can be somewhat more at riskwithout endangering minimal protection.

The larger of the partial privatization plans inthe Advisory Council Report, the PSA plan, in-creases the payroll tax by 1.52 percent and borrowsfrom the federal government ($2 trillion at thepeak), greatly increasing the deficit and the U.S.debt for the next 72 years. This plan would elimi-nate the traditional wage-related Social Securityprogram, substituting flat benefits even though itstill would be supported by a wage tax. The higherpaid would pay many times as much as lower-paidworkers but still get the same benefit.

These proposals do not seem attractive. Aboveall, they weaken the reliability of the nation’s basicretirement system by shifting it in part from adefined benefit to a defined contribution system.This approach requires workers to set aside addi-tional income for the sole purpose of retirement,regardless of what families may require for protec-tion against health care costs or to meet otherneeds that are more immediate than retirement.Both plans have particularly harsh benefit cuts forthe totally disabled, and both increase the difficultyof adequately financing Medicare because they takemore from wages for the single purpose of increas-ing retirement savings.

These plans force the first generation or twoof workers to pay twice—once to build up their ownindividual accounts and once to continue paying forbenefits for others under the current pay-as-you-gosystem. The PSA plan does not requireannuitization, and thus it introduces a new uncer-tainty as to whether the accumulated funds willlast an individual’s lifetime. Both plans undoubt-edly exaggerate their contribution to increasingnational savings because they do not take intoaccount any reductions that individuals wouldmake in their current voluntary savings if they arecompelled to save additional amounts through thegovernment program.

■ Administrative ProblemsThe larger privatization plan, the PSA plan, alsopresents quite overwhelming communications andadministrative problems. It would be very difficultto make clear, such as in an informational pam-phlet, the kind of protection people would haveduring the transition period. It is extremelycomplicated. It also is difficult to see just how thegovernment would ensure that 5 percent of wages

93

actually was deducted from workers’ earnings eachpayday and sent to a designated broker, bank, orother financial institution and then kept there, or,if moved to another financial institution, properlyreported again to the government.

This might work for those who now have401(k) plans, but enforcing the provisions of such alaw in the case of hundreds of thousands of smallemployers seems very difficult—much more diffi-cult than the reporting requirements under SocialSecurity. In summary, the six of us found a lot ofproblems in the individual account plans andbelieved that if the rate of benefits and contribu-tions were to be improved for younger workers, wewould have to find another way.

■ Bringing the Current SystemInto Balance

It really isn’t too hard to balance Social Securityover the next 75 years. There are quite traditionalways of doing so, and the six of us who generallysupport the present system have made severalcommon-sense proposals that reduce the2.17 percent of payroll down to 0.80 percent ofpayroll, without major increases in contributions ormajor reductions in benefits. They are listed in theAdvisory Council report, and I repeat that tablehere:

You also can go the rest of the way andeliminate the remaining 0.80 percent of payrolldeficit with changes that are well within thetradition of Social Security’s past changes. You getanother 0.50 percent reduction in the deficit if youare willing to raise the retirement age above age 67as scheduled in present law, and, as a majority ofthe council recommends (although I do not), raise itautomatically in the future in accord with increasesin the length of life. Then, if benefits were cut6 percent (instead of the 3 percent in our plan),these two changes together would bring the 75-yearestimates into balance.

Another possibility is that over the next fewyears the Bureau of Labor Statistics may makefurther corrections in the Consumer Price Index,which governs increases in the cost of living. So itcould be that the program will be brought entirelyinto balance in these ways and without majorbenefit cuts, increases in taxes, or departures frompast practices.

■ Investing ReservesThese changes, however, would not improvemoney’s worth for young workers and would cutbenefits more than we believe desirable. So whatshould we do? We recommend a shift from pay-as-you-go financing to partial reserve financing andconsideration of ways to improve the return on sucha reserve. In particular, we propose a careful andsympathetic study of another way of eliminatingthe 0.80 percent of payroll deficit: Invest a part—upto 40 percent—of the accumulating trust fundsdirectly in passively managed private equities,indexed to the broad market.

The easiest way to get a sense of how muchimprovement in money’s worth results from invest-ing 40 percent in stocks is to study chart 14.1.Money’s worth (that is, the present value of lifetimeprotection furnished compared with the presentvalue of employer and employee payments) differssubstantially for any specific investment scenarioby reason of the composition of the beneficiarygroup and the workers’ ages at the point of timeconsidered. To get a sense of money’s worth for thecombined Social Security population under variousinvestment proposals, the Advisory Council createdthe concept of a “composite worker” and measuredmoney’s worth for this worker by age.

Table 14.1Reducing Social Security’s Long-Term

Deficit

Impact of Specified Changes in Coverage, Calculation of Benefits, Taxationof Benefits, and Adjustment for Correction of the Consumer Price Index

(Shown as a Percentage of Payroll)

Long-Term (75-year) Deficit 2.17

Change Impact on Deficit

Increase taxation of benefits –0.31Redirect taxes on benefits from HI to OASDI (effective 2010-2020) –0.31Extend coverage to new hires in excluded state and local positions –0.22Adjust for Correction of CPI –0.31Increase length of computation period from 35 to 38 years or

increase payroll tax 0.3 percent –0.28Adjust for interaction among proposals* +0.06Total +1.37Remaining deficit –0.80

* The impact of a proposed change may increase or decrease somewhatin interaction with others. For example, taxation of Social Securitybenefits will produce less income if benefits are reduced, and increasingthe wage-averaging period reduces benefits an average of 3 percent.


94

The composite worker is a theoretical con-struct representing single workers, married coupleswith one spouse working, and married couples withboth spouses working in proportion to their pres-ence in the beneficiary population. They areassumed to have lifetime earnings in the properproportion for each demographic group. The SocialSecurity taxes and other payments are the onesappropriate for the birth dates shown in thehorizontal scale. The present value of the SocialSecurity protection furnished takes into accountthe probabilities assumed in the 1995 SocialSecurity Board of Trustees’ report for retirement,disability, and survivors benefits, including eligibil-ity for the benefit rate and the length of payment.The present value of the payments is computed,assuming an ultimate 2.3 percent real valuationinterest rate, or the rate assumed for long-termgovernment bonds, and a 7 percent real valuationrate for investments in the stock market.

The horizontal scale on the chart measuresthe extent to which money’s worth for the particu-lar plan reaches 100 percent of the money’s worthprovided by a fully advance-funded system earningthe long-term bond rate of 2.3 percent. Becausewhat is being measured falls far short of a fullyadvance-funded plan, it is not surprising that themoney’s worth results fall short of 100 percent onthe scale—even though some investments areassumed to earn more than 2.3 percent (7 percentreal, for example, for investments in the stockmarket).

To understand the difference between themaintain benefits (MB) plan with investment in

stock and the present pay-as-you-go plan or the MBplan turned into a pay-as-you-go system after usingup the MB proposals short of stock investment, lookat the three lines on the chart. The MB plan withinvestments of 40 percent in stocks reaches 96percent in the horizontal measure and stays steady(representing a 4.2 percent return on combinedinvestments), as compared with a constant deterio-ration in money’s worth from one cohort to anotherfor each cohort born after 1975 for both pay-as-you-go plans.

The idea of investing part of Social Securityfunds in stocks is a new enough idea for SocialSecurity that we are not urging its immediateenactment. In a program like this, which affectsjust about everyone in the country, we believe thatmajor changes should come about only after thedevelopment of a considerable consensus. So wefavor moving promptly on those common-senseproposals that are relatively easy to do and thatalso will bring the Social Security trust fund muchcloser to balance. At the same time, we hope theAdministration and the Congress will be carefullyexamining and researching the idea of directlyinvesting in the stock market, as is now done byalmost all other pension systems, i.e., practically allstate and local systems and private pensions andmany federal employee plans, such as the definedbenefit plans of the Federal Reserve Board and theTennessee Valley Authority.

We have described a specific plan as a basisfor evaluation of this idea, although, of course, thedetails may well come out differently. Short of agovernment contribution, investment in stocks is

4 0

6 0

8 0

100

120

140

160

1930 1940 1950 1960 1970 1980 1990 2000 2010

MB PL PAYGO MBPAYGO

Bene

fits

as a

Per

cent

age

of C

ontr

ibut

ions

Chart 14.1Money’s Worth Ratios—Total Composite Worker

Comparison of PL PAYGO, MB PAYGO, and MB with Investment in Equities

Year of Birth

95

the only way we envision that can bring the systeminto long-range balance and at the same timeimprove the benefit/contribution ratio for youngerworkers and future generations. If instead benefitsare cut further, or taxes increased more, the resultwill be an improvement in the balance of the planover 75 years; but, at the same time, there would bea worsening—not an improvement—in the benefit/contribution ratios for young workers.

The plan that we have outlined, which limitsthe amount invested in stocks to ultimately about40 percent of the Social Security build-up, neverwould put the government in the position of owningmore than about 2.5 percent to 5 percent of thetotal value of all stocks. Our proposal would followthe general approach of the Federal EmployeesThrift Plan. That is, the President would appointan expert financial board, with members confirmedby the Senate. This board would have only threefunctions:• It would select the index that would govern the

investments in stocks and which would makethe investments representative of the broadmarket.

• It would select, by bid, the portfolio managerswho would keep the stock portfolio in tune withthe index.

• It would report to the Congress and the countryon how the plan was working and also makechanges in portfolio managers and indexes asseemed appropriate.

But these three functions would be the limitof the board’s responsibility. Nobody would bepicking stocks. There is no loss of return in thatprocedure; it is very unusual for active manage-ment over any long period of time to beat marketrates of return. We would not propose that SocialSecurity try to do so.

It also is important to insulate SocialSecurity’s holding of stocks from any influence oncompany policy. So Social Security would not beallowed to vote stock proxies, just as is the case

with the Board of the Federal Employees ThriftPlan. One could eliminate the voting rights out-right or have them exercised, not by the owners,but by the portfolio managers (as is done by thethrift plan). Another possibility is to have thevoting rights of the stocks held by Social Securitycast in the same proportion as the votes cast byother owners.

■ ConclusionI want to make absolutely clear that the six of uswho support the maintain benefits proposal do notconsider the three proposals that came out of theAdvisory Council to be the three best possibleproposals. If our plan were not adopted, we wouldnot select one of the other two as the second-, thethird-, or even the fourth-best plan. We are com-pletely opposed to the idea of partly privatizingSocial Security by reducing Social Security benefitsand putting either new money or part of the SocialSecurity contributions into individual accounts. Wethink there is an overwhelming case against suchplans and oppose them in any guise.

There is no need to make such radicalchanges in the Social Security system. Its financingunder current law is assured until about 2030; andat that time, current law still would support75 percent of the cost of benefits. Even in 75 years,current law would support more than 70 percent ofthe cost of the benefits. The problem is how toassure 100 percent payment from 2030 on. Most ofthe cost of benefits after that date already iscovered. The rest of the job can be done, and doingit is really not that hard.

The situation with Social Security is similarto that of homeowners living in a sound house.They like it very much; they only need to have itsmortgage refinanced. There is no need to tear thehouse down, remodel it, or trade it for a differenthouse. The need is only to improve its long-rangefinancing.


96

97

15The Case for the IndividualAccounts Optionby Lawrence H. Thompson1

■ IntroductionI have agreed to outline the case for the “indepen-dent accounts” approach, the particular optionassociated with the Chairman of the Social SecurityAdvisory Council, Ned Gramlich. I am pleased to doso because I believe it illustrates a general ap-proach that has much to commend it. It is anapproach that attempts to strike a reasonablecompromise among several legitimate, but notnecessarily compatible, objectives of Social Securityreform.

Each of the three options advanced in theAdvisory Council report produces a retirementincome package that on average provides benefitsequal to (or slightly higher than) those scheduledunder present law for steady, long-service workersretiring between now and 2045 or so. Each isadequately financed under reasonable assumptionsabout future demographic and economic trends—not just for the traditional 75-year projection periodbut for the years thereafter. Each raises the valueof the Social Security protection offered to theSocial Security contributions paid for today’syounger workers.

■ Different ApproachesThe proposals achieve these results, however, byemploying different approaches with differentstrengths and weaknesses:

• The “maintain benefits” approach comes theclosest of the three to preserving both theapproximate scope and structure of currentlyscheduled benefits. The financing strategydiffers from the other two strategies in that

some income tax revenues now going to theMedicare are shifted to the cash benefit pro-gram, and up to 40 percent of the assets of thecentral trust fund are invested in equities.Contributions are increased by 1.6 percent in2045.

• The “individual accounts” approach reduces thecurrent benefit package to a level that can befinanced with the current 12.4 percent tax rate,a reduction that averages about 30 percentwhen fully phased in. It adds a new individualaccount financed by a new 1.6 percent employeecontribution. The balances in the individualaccounts are converted into indexed annuitiesand added to Social Security checks at the timethe worker retires. Under the assumptions usedto project the impact of each option, the indi-vidual accounts would, on average, offset thereductions made in the defined benefit portion ofthe Social Security package, producing totalbenefits similar in size to those produced in themaintain benefits approach.

• The “personal security account” approachreplaces the current old-age benefit with a flatbenefit paid at an amount similar to today’sSupplemental Security Income (SSI) benefit.About one-half of the contributions used tofinance today’s old-age benefits, 5.0 percentagepoints, are diverted to individual accounts. Theindividual accounts are managed by a privatefirm chosen by the individual worker, subjectonly to minimal regulation about how they areinvested. Workers are free to withdraw accumu-lated assets however they like after qualifyingfor retirement. The revenues lost by diverting5.0 percentage points to individual accounts arereplaced by massive new government borrowing;the new debt is retired over a period of 72 yearswith the proceeds of a 1.5 percent payroll tax.

1 Senior Fellow, The Urban Institute. The viewsexpressed here are those of the author and not necessar-ily the Urban Institute.


98

I see two important advantages in the indi-vidual accounts approach:

• The approach responds to the desire of manyyounger workers to incorporate a element ofindividual accounts into the Social Securitypackage; however, it does so while preservingthe advantages of having a substantial definedbenefit component to the retirement incomebenefit package and the administrative efficien-cies inherent in a coordinated system for collect-ing and managing funds.

• The approach encourages increased personalsaving that will help increase the size of theeconomy from which future retirement benefitswill have to be paid. It does so, however, withoutsacrificing the effectiveness of our Social Secu-rity program as a mechanism for assuringadequate retirement incomes or relying on asomewhat problematical mechanism for financ-ing the transition to greater advance funding.

■ Providing for IndividualAccounts

As its name implies, the “individual accounts”approach provides each worker with an individualaccount as an integral portion of the Social Securitypackage. The account is financed through a manda-tory contribution. To the worker, the system looksvery much like many of today’s 401(k) plans. Theaccounts are managed centrally, but workers areoffered several options as to how they wish to havetheir accounts invested. These might includeseveral broad-based stock index funds, several bondfunds, and a couple of money market funds. Work-ers could participate in their own retirementplanning by selecting the investment package theyprefer. When they reach retirement age, theamount in their individual account would beconverted into a price-indexed, life annuity andpaid to them as an integral part of their SocialSecurity benefit.

One advantage of this approach is that itintroduces individual choice in investment vehicleswhile helping to insulate workers from some of therisks inherent in retirement planning. The insula-tion comes both through the preservation of a majorrole for a defined benefit component of Social

Security and through the requirement that indi-vidual account balances be converted into indexedannuities.

■ Risks Associated WithRetirement Planning

In principle, workers could save for their ownretirement simply by setting aside a fixed percent-age of their earnings each pay period without anyneed for a government program. Far-sightedworkers entering the labor force who decided toplan ahead for retirement quickly would discover,however, that their planning needed to deal withseveral risks inherent in retirement planning.

Cohort Life Expectancy

First, to know how much to set aside each payperiod, workers will need to know how long theircohort can expect, on average, to live after reachingwhatever retirement age they select. This requiresprojecting mortality trends some 40 to 60 years intothe future. Such projections are currently a matterof heated debate among demographers and one ofthe areas of greatest uncertainty in SocialSecurity’s long-range projections. If mortalityimproves more than workers expect, they will findthat they have not saved enough when they reachretirement age. If mortality does not improve asmuch as they expect, they will have saved toomuch.

Future Economic Growth and InvestmentReturns

Faster economic growth implies a more rapidincrease in living standards over the course of theworkers’ lives, leading to the need for additionalresources if preretirement incomes are to bemaintained. The difficulties in projecting invest-ment returns is one of the key topics that theEmployment Benefit Research Institute (EBRI)hopes to research over the next few years. Thechallenge is to do the best job possible in under-standing the implications of what inherently isuncertain.

Relative Earnings Level

A third source of uncertainty is the earnings levelof each worker relative to the entire cohort. In

99

general, those with lower earnings will find savingfor retirement more difficult, but they also willneed to have a higher benefit (relative to previousearnings) to preserve their preretirement livingstandards. In a world of corporate downsizing,workers who may have thought that they couldcount on high incomes and generous employerbenefit packages may find their expectations upsetmidway through their working careers.

Individual Life Expectancy

Once the cohort reaches retirement age, eachmember must deal with the uncertainty over his orher own life span relative to that of the othermembers of the cohort. Those who die relativelysoon after reaching retirement age probably willhave saved more than they needed to and will leaveestates. Those who live beyond the average lifeexpectancy of their cohort run the risk of outlivingtheir assets.

Inflation

Finally, workers entering retirement face theuncertainties associated with future changes inprice levels. Inflation averaging 7 percent annuallywill reduce the purchasing power of a givenmonthly benefit by one-half in 10 years and bythree-quarters in 20 years.

■ How Risks Are Distributed inDifferent Retirement IncomeVehicles

Different retirement income institutions handlethese risks in different ways. The old-age benefitportion of the current Social Security program is adefined benefit plan designed so that society as awhole shares each of these risks with the individualworker. It is precisely this feature of Social Securitythat makes it social insurance.

Employment-based defined benefit pensionplans also help workers deal with several of theserisks. In particular, the plan sponsor assumes therisks associated with changes in cohort life expect-ancy, uncertain investment returns and rates ofeconomic growth, and (in most cases) the risk thatany given individual will live beyond his or hercohort’s life expectancy. Employer-based plans donot deal as well with uncertainties about the

individual’s earnings relative to the cohort averageand tend not to deal with inflation after retirement.

In most defined contribution plans, workersbear all five of these risks by themselves. Theworker can reduce the fourth risk, that of outlivingone’s assets, through the purchase of an annuity,but the market for individual annuities is veryimperfect. Workers easily may lose a quarter of thevalue of their accumulated assets in the process ofconverting them into annuities if done on anindividual basis. In principle, in the future theworker also will be able to deal with the fifth risk,post-retirement inflation, by purchasing a govern-ment bond indexed to the consumer price index. Ineffect, the government will assume that risk for anybond purchaser, just as it now assumes the risk forSocial Security beneficiaries.

■ Preserving a Mixed SystemThe current retirement income system is a mixedsystem in which the Social Security program isdesigned to provide about one-half (more for lowerearners and less for higher earners) of the incomeneeded to preserve one’s preretirement livingstandards. The balance must be supplied byemployment-based programs or through personalsavings.

Unfortunately, the trend among employment-based programs is toward increased reliance ondefined contribution systems and reduced relianceon defined benefit systems. Thus, the employer-provided portion of the system already is moving inthe direction of forcing individual workers toassume a greater portion of the risks inherent inplanning for retirement. Although there is no magicformula for dividing these risks, it would seemunwise to convert to a system that transfersvirtually all of the risks to the individual worker.

The individual accounts option allows risk tocontinue to be shared between the worker andsociety. It preserves a substantial role for thedefined benefit portion of Social Security, and itconverts the balances accumulated in workers’accounts into price-indexed, life annuities.

■ Administrative EfficiencyThe individual accounts approach piggy-backs onthe same mechanisms currently employed to collectSocial Security contributions. This is a very effi


100

cient mechanism for transferring and recordingworker contributions. It probably is the onlymechanism that makes sense for a mass programlike Social Security.

Each year, the Social Security Administrationposts some 190 million different wage items. Eachitem represents the annual earnings of one workerderived from one employer. The median wage itemis $15,000. A worker who earns $15,000 over thecourse of a year receives, on average, $575 everytwo weeks. A worker required to contribute5 percent of his or her earnings to a mandatorysavings account (the proposal underlying the PSAapproach), would be making a contribution of justunder $30 out of each pay check.

Now, imagine a world in which every workergets to select a mutual fund company to manage hisor her assets and in which each worker’s employeris responsible for withholding and remitting theregular contributions required of this worker. Theemployer of a modest-sized work force could easilyfind that every worker had selected a differentmutual fund company, leaving the employer to sendeach such company a check in the amount of $30(more or less) every two weeks.

In addition to the employers’ costs, theindividual mutual funds will incur operating andmanagement costs, which the Advisory Councilassumed would amount to about 1 percent of theassets being managed. While indexed mutual fundsclearly can manage moderately large accounts atthese or lower cost levels, it is not clear howcheaply they can manage the much larger numberof much smaller accounts this plan entails. Nor canmore actively managed or specialty funds operateat these expense levels.

Surely, there must be a more efficient way torun a series of individual accounts than to have190 million checks with an average face value of$30 running through the banking system every twoweeks and to have hundreds of different firmsduplicating administrative structures to managemillions of modest-sized accounts. There is: theindividual accounts approach, under which thepayment system is centralized so that each em-ployer needs write only one check—to cover incometax withholding, regular Social Security contribu-tions, and the mandatory individual supplementalcontribution—and under which fund managementis streamlined.

As compared to the other approach to indi-

vidual accounts, workers also will benefit from:

• lower administrative costs associated withmanaging their money,

• lower administrative costs associated withmaking regular benefit payments, and

• lower transactions costs associated with convert-ing account balances into annuities.

■ Increased Personal SavingsMoney accumulated in the individual accounts willbe invested in private-sector securities. It willrepresent additional funds available to financedomestic investment. To some degree, this shouldfacilitate an increased pace of domestic investment,leading to somewhat faster economic growth.Additionally, to some degree, it may displace thefunds that now are being imported from abroad tofinance U.S. domestic investment. In either case,our national income would be higher in futureyears.

All three options contained in the AdvisoryCouncil report call for increased investment inprivate-sector securities as a key element inincreasing the benefits associated with a given levelof worker contributions. Under the individualaccounts and personal security accounts options,such investment also is explicitly designed toincrease national saving. Of the two, however, Ibelieve that the individual accounts is the far morereliable approach.

The individual accounts approach follows avery straightforward strategy: Workers are askedto contribute an additional 1.6 percent of theirtaxable earnings to their accounts. This is an add-on to the current Social Security tax. It forces areduction in consumption, which is the only way inwhich savings can be increased.

The personal security account approachimposes an additional payroll tax of essentially thesame size. If the plan actually were to be executedas designed, it would have a very similar impact onnational saving because it essentially would havethe same impact on aggregate consumption.However, the plan relies on a much more complexfinancing arrangement that easily could turn out tobe far less effective. The additional payroll tax doesnot flow directly into additional saving; it is used toretire a huge special issue of government debt that

101

is issued, in turn, to finance the transition betweenthe current system and the system envisioned bythe proponents of the personal security option. Theplan rests on the somewhat risky assumption thatfuture Congresses will consistently decide tocontinue levying a payroll tax on working Ameri-cans for the sole purpose of retiring debt, a tax thatwould seem to be an obvious target in times ofeconomic distress or political competition.

■ Preserving Medicare FinancingOne troubling aspect of the maintain benefits planis the diversion to the retirement and survivorsprogram of some of the resources currently goinginto Medicare. In effect, such diversion amounts toa hidden tax increase because it creates the need tofind substitute resources for Medicare. The indi-vidual accounts plan avoids such financing devices.

■ ConclusionOver the next half-century, demographic shifts willforce an increase in the cost of supporting the agedpopulation. One implication is that currentlyscheduled Social Security benefits cannot befinanced with currently scheduled revenues.

The members of the Social Security AdvisoryCouncil concluded that the system ought to pre-serve a set of benefits which, in the aggregate, weresimilar in scope to those scheduled currently. Eachalso came to the conclusion that one way to financethese benefits was through some form of equity

investment in the Social Security program.A benefit package fairly close to the current

package can be financed directly if some of themoney now scheduled to go to Medicare is divertedto the pension program and up to 40 percent of thereserves of the program are invested in equities.Such a package does the best job of preserving thecurrent defined benefit protection for futureretirees, but it makes no provision for individual-ized accounts.

A radically different approach involves cuttingby some 60 percent the current defined benefitcomponent of the program and shifting the majorityof the responsibility for future Social Securityretirement benefits to relatively unregulatedindividual accounts. This approach does introduceindividual accounts, but it sharply reduces thedegree of worker security provided by definedbenefits paid as real annuities.

The individual accounts approach representsa reasonable compromise between these twoapproaches. It does require an immediate increasein the total amount being contributed for SocialSecurity, which is precisely why it has the potentialto increase national saving. It also allows for adegree of self direction through a system of indi-vidual accounts; but it preserves 70 percent of thecurrent defined benefit package and forces theadditional resources accumulated in the individualaccounts also to be annuitized. This approachstrikes a reasonable balance between potentialimprovements in the macro economy and mainte-nance of adequate worker benefits.


102

103

16Comments on the National ThriftPlanby Neil Howe

■ IntroductionThe assertion was made several times that the factthat individuals can’t beat the broad market is anargument against personal accounts. And obviously,those of us who are proponents of this approach sayjust the opposite. That’s what makes personalaccounts possible. It means that people can investin indexes, and they can invest in broad marketinstruments, and do just as well over the long termas those who pay commissions to a lot of fancybrokers.

What people really want assured under thissystem is that, when they invest in a personalaccount, not that they can get some broker to makeclever choices but that, first, their money will beeconomically saved. Most people think that is trueabout Social Security, and when you inform themthat it is not, there is genuine surprise. They areshocked—particularly older people, not so much theyoung. They know basically how the fiscal account-ing works.

The second consideration is that it is moneyover which they are assured ownership, and this isa basic political and constitutional point. I thinkthat, particularly after I have talked to a lot ofyoung people, what they are really concerned withis the ability of the government to keep promises,perhaps even more than the ability of the DowJones to live up to its expectations.

■ The National Thrift PlanLet me start by explaining how the National ThriftPlan came about. Setting out, we had two overallobjectives, the combination of which put us some-where in the middle of the Social Security reformdebate. Or at least it put us somewhere in themiddle among those who were looking at reform.

On the one hand, we wanted to design avisionary plan that would truly go after the root

problems of the current Social Security system: itsmassive unfunded liabilities; its enormous pro-jected fiscal deficits; its disincentives to work effortand private-sector savings; its evaporating level ofpublic trust, particularly among the young; and,most seriously, a low and declining rate of return onSocial Security contributions. This definitely put uson the visionary side of those who simply want afew reforms, as little as possible, to keep the systemwe have.

Specifically, we wanted a fully funded planthat would allow workers to invest their contribu-tions into the real economy, and a plan that, oncemature, could never run a deficit. We wanted theassets of this plan to be personally owned andmanaged as common law property of each contrib-uting worker. Unlike Social Security, these retire-ment assets could never be changed or reduced byCongress, nor could government ever use them topaper over its own deficits, its own red ink.

We wanted this plan to afford a more gener-ous protection for elder Americans in poverty thanour current system. We also wanted it, like ourcurrent system, to subsidize the retirement securityof low-income workers, but to do so on the savingsside before retirement, not on the benefits sideafter retirement. That is the whole concept of thematch that we provide for low-income individuals,basically using the Social Security Administrationdefinition of low-income workers, which is45 percent of the average wage.

We wanted the plan to offer all new retireeshigher total benefits, and eventually to offereveryone a much better retirement deal on theircontributions than the current system provides.Most importantly, we wanted the plan to boostproductivity and real wages throughout theeconomy by adding hugely to the national savings. Iam sure if you look at our plan you will see thatthis effect on national savings is quite profound.


104

We wanted a visionary plan. Yet, on the otherhand, we also wanted to design a realistic plan. Wewanted a plan that is not predicated on newgovernment debt, or new taxes, or unrelated outlaysavings, and we wanted a plan that does not minda government program that guarantees a safetynet, and requires people to do something to preparefor their own future. That is what a public sector isfor, and that is why we actually avoided the term“privatization,” which we think is somewhat of amisnomer for this whole concept. This puts us onthe realistic side of those who dream about utopia.

Specifically, we wanted to base the plan onvery conservative assumptions about the likelyresponse of our economy. We wanted the plan toavoid any financial slight of hand, such as havingthe government play the spread between thereturns in debt and the returns on equity. Wewanted to avoid assuming that magical gains ineconomic efficiency would suddenly pay our way.We also wanted to avoid shunting the transitioncosts into the future onto today’s younger genera-tions, or onto unborn generations, or onto someother part of government.

In fact, we wanted the plan to guarantee thatthe transition costs would require neither anincrease in any year in total federal debt obliga-tions nor any increase in any public fee or taxunrelated to the Social Security system.

And finally, we wanted the plan to pass thecriterion of what might be called interruptibility,meaning essentially that starting this plan wouldmove us, socially and economically, in a desirabledirection, even if we had to stop the plan before itwas fully implemented. Much of the discussion Ihave seen of Social Security reform does not takethis into account. In fact, no plan having to do withSocial Security, including the original 1935 act, orthe 1939 act, and perhaps not the 1983 act, hasever come fully to fruition as the people whodesigned it planned. History changes. Politicschange. All kinds of things change as the worldmoves on.

What is very important to understand is that,rather than, for instance, putting into stone theunfunded obligations of today’s young, beforestarting to pay them off in the future, we shouldmake sure that we aren’t digging a deeper holebefore we start climbing out of it. And one of thethings that our plan does is increase national

savings in every year and also decrease the un-funded benefit obligations in every year. It puts oureconomy on a more fully funded or investablefooting than otherwise.

■ Questions about ModelingLet me make a couple of other points concerningthe modeling question. How do you model SocialSecurity reform? What kinds of modeling questionsare important here?

I think the interruptibility question is vital.Another thing that we have been very concernedabout in looking at other plans is what I call theceteris paribus standard: All other things beingequal, we need a standard by which we can com-pare one reform plan to another, and what thatmeans is that you can’t simply say, “Well, thenthere are the transition costs,” right? I have anideal plan, and then there are five generations thathave to get there, but we don’t want to talk aboutthat. You have to be able to describe very explicitlywho is going to absorb these costs in a way in whichthese different plans can truly be compared. Wehave tried to do that totally within our system. Wedon’t try to do anything outside the system.

Finally, the savings, productivity, real wageslink is absolutely essential. It is this link thatkeeps Social Security reform from turning intosomething close to a zero sum game. In fact, as I’vebeen pointing out occasionally to the EmployeeBenefit Research Institute, it even changes thewhole meaning of the benefit/payback ratios.

Think about it. If productivity starts improv-ing—say significantly—5 to 10 years after a systemlike this is implemented, 30 or 40 years from now,productivity or real wages will be quite signifi-cantly higher. This means the same labor effort inSocial Security contributions is going to produce amuch greater payback, because it’s going to beworth more in dollars.

This is the money-back twist that is veryoften not considered in the way these calculationsare done. It is easy to do money-back calculations ifyou consider that productivity is going to be thesame in all the various scenarios. But once youstart considering an entirely different economictrajectory, particularly after 40, 50, 60 years, youhave to revise your whole thinking about paybackratios.

105

■ ConclusionSusan Dentzer1 asked why we discuss SocialSecurity rather than Medicare? After all, Medicarehas future liabilities that are three times as large.Thus, could it be because, in Social Security, there’sa killing to be made by certain Wall Street firms,and so forth.

Personally, I would absolutely love to talkabout Medicare at the same time as Social Security.My problem throughout my history of involvementin these issues is that I was always told, by peoplewho didn’t really want to do much, that SocialSecurity and Medicare were two entirely differentproblems, and we could never talk about themtogether. In fact, all of these different entitlementproblems were like so many hardened silos, if youknow what I mean.1 See Susan Dentzer, “Social Security Reform: Gaps in

Perception,” in this volume.


106

107

17A Retirement System for theFuture: Personal Security Accountsby Ann L. Combs

■ IntroductionI would like to discuss my reasons for thinking thatmoving in the direction of a partially privatizedSocial Security is the correct direction for thecountry and merits serious thought and attention.

I do not begin to deceive myself that we havethought through, or solved, all of the issues and theproblems. They are very real, and I think themodeling that Martin Holmer and Jack VanDerheiare doing with the Employee Benefit ResearchInstitute is really essential, because partialprivatization of Social Security would represent afundamental change in the way we deliver retire-ment income, and we need to work together toanswer all these questions. But it certainly is anissue that is worth debate and discussion.

When the Advisory Council on Social Securitybegan our deliberations, we were faced with adilemma. There are no easy choices; there is nosimple solution, or painless option. We are facedwith a program that is about 25 percent to30 percent underfunded over the long term. Therewas a strong reluctance to raising payroll taxes,both for economic reasons on the part of severalmembers and for political, realpolitik reasons onthe part of others.

We started in 1994, a long time ago. At thetime the Council was considering payroll taxoptions, no one believed that the report would betaken seriously if substantial payroll tax increaseswere a key element.

I am referring to substantial payroll taxincreases that would be imposed to continue payingbenefits under the current system. A distinctionshould be drawn between taxes that fund thecurrent pay-as-you-go system and taxes thatfinance a transition to a new, substantially funded,system.

In addition to opposition to payroll taxes,there was also legitimate and grave concern about

the kinds of benefit cuts that would be necessary tosolve the equation without any additional revenuecoming into the system. We all shared thoseconcerns. As a result, the Council started talkingabout the whole notion of investment in the privatesector as a way to increase the trust funds’ invest-ment return, as opposed to draconian cuts inbenefits or raising payroll taxes.

Five members of the Council support thepersonal savings account (PSA) approach. Frankly,I personally backed into supporting it. Once westarted talking about investments in the privatesector as a way to solve this dilemma, I simplycould not get comfortable with the notion of central-ized investment by the federal government in theprivate markets, and so, firmly believed that themore we can decentralize that aspect and put itunder individual control, the better it would befrom an economic perspective, and from a politicalperspective.

The PSA approach was developed for veryvalid policy reasons.

■ The Personal Savings AccountPlan

The PSA plan works as follows. We would gradu-ally transform the current Social Security systeminto a two-tier system. The first tier would be a flatdollar benefit that would go to all workers, andwould be financed on a pay-as-you-go basis. Thesecond tier would be an individual account, a PSA,financed with 5 percent of the current payrolltaxes. It is a reallocation of the current payroll tax.

Workers would own their individual PSAs.They would be able to invest in a wide range offinancial instruments of their own choice. Theycould withdraw assets from the PSA only after age62. I agree with all the comments about the dangerof people spending down their account balances if


108

early access is allowed, and I recognize thepolitical difficulty of walling this money off. Ibelieve it is essential that the assets in a PSA bepreserved for retirement. This is an importantissue that requires further discussion. Theseaccounts would be an essential part of SocialSecurity. They are different from individual retire-ment accounts (IRAs). PSAs should be inviolateuntil age 62.

In addition, our plan does not requireannuitization of the account balances. Annuiti-zation is an option, but it is not required. As aresult, any balance remaining upon death couldpass to one’s estate.

Another thing that is important to keep inmind is that the new system would not affectanyone over age 55. It would be phased in verygradually for people between age 25 and age 55,and only those under age 25 would be fully underthe new system.

We make some changes that apply across-the-board: raising the retirement age, repealing theretirement earnings test, and changing the taxa-tion of benefits. But the fundamental restructuringof the system is phased in slowly, and everyone overage 55 is grandfathered in the current structure.

■ The Flat BenefitThe advantage of the first tier of our plan—the flatbenefit—is that it is the safety net. It provides thebasic floor of protection. It is an indexed annuity. Itredistributes benefits, very substantially, fromhigh-paid to low-paid workers, and I think that isan important feature of the Social Security systemthat we should not lose.

And in some sense, it shields low-wageworkers from some of the risk of the private-sectorinvestment, because more of their retirementincome, relatively speaking, would come from theflat benefit. High-wage workers are going to bemore dependent on their individual accounts, whichis probably appropriate, considering the risk.

That being said, I want to comment on thespecifics of the flat benefit. The dollar amount ofthe benefit has been criticized as being inadequate.In the proposal, it is $410.00 a month, which isabout two-thirds of the poverty level. That is a lowbenefit. It is what we could afford to do with the7.4 percent of the payroll tax that is being used to

finance both that flat benefit and the disabilitybenefits. It also is designed to provide a safety net,supplemented by the account balances in anindividual’s PSA.

It needs to be made clear who would receivethe flat dollar benefit. First of all, no one over age55 would be affected. As I have said, they would begrandfathered. People between age 25 and age 55would receive a blended first-tier benefit. Theywould receive a portion of what they are entitled tounder current law, based on how long they haveworked under this system, and a portion of the flatdollar benefit.

While the system is being phased in, the flatdollar benefit would be indexed for growth inwages, so that by the time the people who are nowunder age 25, those who would be fully under thenew system, retire, that benefit is projected to beequal to $629.00 in 1996 dollars. This is becausereal wage growth is expected to outstrip inflation.Thus, the present value of the benefit would growover time. Also, once an individual retires andbegins drawing benefits, the flat benefit would beprice-indexed to reflect the cost-of-living, just asSocial Security is today.

Moving on: what are the benefits of theindividual account portion of the PSA plan? Themost fundamental is that it would replace a systemof unfunded benefit promises that is using currentworkers’ taxes to pay for current retirees with asystem in which each generation is responsible forsaving for its own retirement.

It is fundamentally different from either ofthe other two approaches recommended by theCouncil. At the end of the day, under the PSA plan,we end up with a benefit that is 50 percent fundedfor each worker. I think it creates a very direct linkbetween the taxes that are paid and the benefitsthat are received, eliminating some of the complex-ity of the current benefit formula, making it moretransparent so that people understand what isgoing on, and also eliminating some of the labormarket distortions.

This approach would minimize the sensitivityof a significant piece of the system to demographicchanges, because participants are not going to bedependent on future workers, and how many ofthem there are, to support them in their retire-ment. It reduces the current system’s incentives toretire even though you may still be productive. The

109

defined contribution portion of the benefit, in otherwords, is neutral with respect to the decision toretire.

It would allow individuals and their familiesto be more involved in a financial decision thataffects a critical part of their lives. This could havesome secondary effects. I think the presence ofindividual accounts would serve to increase finan-cial literacy, which is good for the economy as awhole. It could encourage additional savings. Ifpeople can see what compound interest does, itmight also encourage those with small savingsamounts to start saving on their own. Many peopledon’t have 401(k) plans, and this might provide amethod for them to see the value of saving, andperhaps to supplement PSAs with additionalsavings. There is a question of whether there wouldbe substitution; people would put less money intheir 401(k)s or other savings vehicles, because nowthey have to put it in a PSA. This substitutioneffect would be minimized under approaches suchas ours, however, because we are reallocatingexisting payroll taxes rather than imposing newmandatory savings.

We have talked a lot about economic growth. Ibelieve that increasing savings—real savings in theeconomy—can improve economic growth, and that,in fact, is the way to deal with the entitlementproblem.

■ Transition CostsSome of the criticism of our proposal concerns anissue that reminds me of the elephant in the middleof the room that everybody’s trying to ignore: thetransition costs of moving to a privatized system.People have asked, “What is the difference betweena 1.52 percent payroll tax to finance the transitionand an additional 1.6 percent payroll tax to fund anindividual account?” The answer is that it matterswhere you end up. Under the PSA plan, therewould be a much smaller pay-as-you-go system andfully funded defined contribution plan accountsthat participants own. The individual account planresults in a very large unfunded pay-as-you-goprogram financed with 12.4 percent of payroll andsmall individual accounts.

Our transition tax is not a new payroll taxgoing to fund the current system and the currentlevel of benefits. It is a very different system; and

that is a distinction about which people need to beeducated. It is very important.

I am proud of the fact that we at least tried toaddress the transition problem. I think people havetried to hide the pea a little bit in this debate, andthat is not honest. The transition costs are real andthey are very substantial. And, as you know, itoccurs because some generations are going to beasked to start paying for their own retirement inaddition to continuing to pay for current retirees.

According to Social Security, the system isprojected, over the next 75 years, to pay $18.6trillion in benefits. In contrast, the current reserveand income taxes are equal to $9.8 trillion. In otherwords, you can think of it as an unfunded liabilityof $8.8 trillion. What we are doing is taking thatimplicit federal debt—and it is a debt we owe—andmaking it explicit. The 1.52 percent is the equiva-lent payroll tax increase that the actuaries esti-mate it would take to retire that debt.

We have tried to finance the transition with acombination of payroll taxes and federal debt. Thedebt is issued to smooth out the cash flow, becausethe costs are much higher in the early years duringthe baby boom generation’s retirement and drop inlater years. The debt is used to smooth out thepayments. We finance the transition over twogenerations, over 70 years. Frankly, all the mem-bers of the panel would prefer to see a consumptiontax, or some other combination of reduced federalspending, as opposed to payroll taxes. But forpurposes of analysis, that is what we went with.

It is important to note that these are esti-mated costs, and worst-case estimates at that. Theydo not assume any economic growth as a result ofany of these savings. They also do not assume anysecondary economic feedbacks. They are justestimates. We do not know for sure that1.52 percent of payroll will be necessary to fund thetransition.

■ RiskOn the issue of risk, I don’t accept the notion thatworkers are not capable of investing for their ownretirements. I do believe we need a much greatereducation effort, and we need to give people thetools they need to learn.

I think the IRA/401(k) experience has beenpositive. It is important to remember, again, theindividual accounts in our proposal are only


110

available to people under age 55. There is alot of the anecdotal evidence about “my mother, mygrandmother.” I would not want my mother to tryto invest on her own at this point, either. However,I think most people under age 55 have had invest-ment experience through their employers, andothers will rise to the occasion.

There is a wealth of information available. Ithink the interest of the financial markets in thisarea will have a good effect, in that they will beinspired to bring more information to bear to makethis work. On the administrative costs issues, I alsothink that Wall Street will contribute a great dealto the debate. They have been very creative insolving problems with respect to multiple accountsand small dollar amounts in other areas, and canhelp develop solutions to those challenges in thisarena as well.

I also believe giving individuals real owner-ship rights is important. We should not forget thatthere is a vulnerability in the current system, apolitical vulnerability, over which people have nocontrol. There is risk in investing your own money.There is also risk in the current system. We knowbenefits are going to be cut, or taxes are going to beincreased. The status quo—the true status quo—isnot an option. And that is a risk over which peoplehave no control and for which they cannot plan. SoI don’t think it is fair to say that there is no risk inthe current system.

■ The Government as InvestorConcerning the maintenance of benefits approach, Ihave fundamental problems with the government

directly investing in the private markets. I wouldpoint out that all of the analysis in the Council’sreport is premised on the fact that, eventually, upto 40 percent of the trust fund would be invested ina private passive equity fund. Eventually, thegovernment would control $1 trillion of SocialSecurity assets invested in the private sector.

I appreciate the efforts that proponents of thisapproach have made to build fire walls and talkabout fiduciary duties, but in the end it is politi-cians who make the decisions and who can rewritethose rules. I believe the temptation for socialinvesting, or targeted investing, or whatever youwant to call it, would be too great, whether it isdirect investment or actively managing the indexthrough decisions to include or exclude certaincompanies from the “passive” index.

I also think there are inherent conflicts ofinterest with the government investing in regu-lated industries. I think the corporate governanceproblems are very real, and cannot be avoided. Ithink government investment would introduceanother element that would take away from fiscaldiscipline, because it would be too tempting forCongress to fiddle with the rate of return assump-tions, or the asset allocation model, rather thandeal with the difficult issues we face in designing aretirement system for future generations.

■ ConclusionI believe that moving toward a system of individualaccounts, where individuals have true ownership,true control and responsibility, is the superiordirection.

111

18 The Case for Incremental Changeby Stanford G. Ross

kinds of ways to look at this, but one way is simplyto look at the average replacement rate now used infuture calculations. The average replacement ratein Social Security today is roughly 42 percent.Under current law, without any changes, that ratewill decline by 2030 to about 36 percent, mainlydue to the raising of the retirement age to 67. Andthen if you try to remedy the 2.15 percent shortfall,you will end up with a Social Security system inbalance, with roughly a 30 percent replacementrate.

Within that parameter you might make itmore efficient. You might make sure that you trulyare protecting lower-income workers, but I wouldgo ahead and take the changes that the “maintainbenefits group” had, plus some more, and try to do,in effect what Chairman Edward Gramlich’s plantried to do: bring Social Security into balancewithin the existing payroll tax.

I do not favor the proposal to invest thereserves in the equity markets because that type ofdirect investment introduces all sorts of politicalrisks into the system. Even countries as committedto public pensions and democracy as Sweden andJapan have had political difficulties with that kindof approach. Further, even if one Congress might doit right, another Congress might do it wrong. I alsodisagree with those who believe that a direct-investment approach will appeal to youngerworkers. Their involvement is too remote; theywould see it as the government doing the saving,not themselves. It would look to them very muchlike the government bonds that are there now.

■ A Case for the VoluntaryApproach

So how should we begin? I would urge consider-ation of a voluntary approach. I would begin toundertake the massive public education that is

■ IntroductionThree issues are very much in the public’s mind:Medicare, tax reform, and Social Security. Thelatter probably would be in third place in terms ofthe likelihood of attracting presidential and con-gressional attention in 1997. Medicare poses animmediate and huge crisis. Tax reform has been afestering issue, as shown by the ClintonAdministration’s targeted tax cuts, Republicanpresidential candidate Robert Dole’s 15 percentacross-the-board cut, and the tax proposals withinthe Republican Party’s Contract with America. Weneed to remember that the way in which Medicareand tax reform issues develop may well influencethe way we look at the issue of Social Security.

If one assumes that any Medicare solutionwill absorb any available resources (and indeed,Social Security will be lucky, if the past is prologue,not to lose some of its payroll tax support to solvingMedicare’s problems), and if one assumes that taxreforms will not undermine the employer deductionfor health care or pension contributions, so that youpretty much can deal with Social Security the wayit looks now—which I think is a big premise—thenthe question is, what ought you to do? I do notbelieve that a program that basically has operatedreasonably well over 60 years needs to be radicallychanged; it needs to be incrementally changed. Butit needs to be changed because it should be broughtinto financial balance, and then it is hoped that alot of the doomsday rhetoric can disappear.

Particularly among younger age groups, Ibelieve that there is a desire for a more individual-ized element in the program. They are less accept-ing of government paternalism. In addition to that,you somehow have to do a number of things at onceif you want to strengthen Social Security.

If you take the 12.4 percent payroll tax, andhope that you can maintain that for Social Security,what does that mean for the future? There are all


112

required by creating a program to include institu-tionalized education. For example, on a voluntarybasis, people could put, say, 2 percent, with taxincentives, into an individual account that would bean add-on to Social Security.

The bill sponsored by Sens. Alan Simpson(R-WY) and Robert Kerrey (D-NE) had somethinglike this in its elective feature for the 2 percent.The problem with that bill was that when you gotthe 2 percent out of the existing 12.4 percent, theremaining 10.4 percent did not leave enough roomfor a Social Security program that maintained thetraditional values and structure of the program.And for that reason, I do not believe that theremaining Social Security program under the5 percent approach that was described leaves aSocial Security program that would have popularsupport or long-term viability. But I believe youcould have a viable Social Security program withinthe full 12.4 percent. If the public were to decidethey wanted a larger program, taxes would have tobe increased at that time to pay for it.

Voluntary worker participation would demon-strate the desire for savings. The program shouldbe made coterminous with Social Security. Unlikeexisting individual retirement accounts (IRAs) and401(k)s, there should be no option to withdrawsavings early for education, medical emergency, etc.It should be available only when Social Securitybenefits are available. So it would be savings lockedup strictly for retirement.

This approach allows people to vote with theirpocketbooks, if you will. Those in the youngergenerations who want to save more can save more,and those who do not will have passed up theopportunity but still will have Social Security. Inshort, I believe we ought to be looking more at avoluntary approach to develop an additional pillarof savings. In that way, the basic Social Securityprogram remains in place for those who invest theirmoney imprudently or are unfortunate. Similarly,you might need a lot less government regulationthan would be required with a mandatory program.As with IRAs and 401(k)s, some restrictions

probably would have to apply, but there would be alot fewer restrictions than if it were a mandatoryprogram.

I also have considered providing some greaterincentives by perhaps including an employer-matching element or some sort of matching out ofexisting tax sources. Lest you think I am a back-door spender by proposing these tax incentives, Iassure you that there are plenty of existing taxexpenditures we could recapture and use to pay forthis purpose. So this would have no net impact onthe budget.

I have not had the time that the three groupson the Advisory Council have had to work throughthis “add-on” voluntary approach, but I think thatit has some appeal because it gets you started in adirection that meets with a lot of enthusiasm. Andthat enthusiasm is not restricted to Wall Streetpeople who create and sell products. I believeyounger people are interested in these issues in adifferent kind of way than their elders, and thiscould be a vehicle for encouraging younger peopleto increase savings.

Any changes here are going to have to bebroadly bipartisan and based on a great deal ofpublic education. Different people make differentcalls on the economics and the politics and howthey think people react to things. One advantage ofa voluntary approach is that it would elicit valuablefeedback about how the people who would beaffected feel about putting away more for theirretirement.

■ ConclusionEven the best models in the world will not give usthe kind of answers we want, because no matterhow detailed and complicated you make them, theynever will fully mirror the way people will react.That is why it would be wise to have a programthat began to provide some real-life experience withthese very difficult issues of increasing savings forretirement.

113

Keeping Track of Social SecurityReform Proposals: A Summaryby Kelly Olsen

Appendix

A■ IntroductionAmong many policymakers across the politicalspectrum, there seem to be two main points ofconsensus about Social Security. The first is thenecessity for reform. The second is that reformought to involve private market investment ofretirement taxes1 and/or of national definedcontribution-style savings in order to give taxpay-ers a better return on their money than the currentsystem affords and thereby reduce national retire-ment program costs. Here agreement stops anddebate about specific reform packages begins.

The Social Security Trustees stated in their1996 annual report that, under intermediateassumptions about the economy and demographics,Social Security will be unable to meet its promisesto beneficiaries by 2029. At this date, the trust fundsurplus will be exhausted, and FICA tax revenuealone will be able to provide only 76 percent ofbenefits payable.2

Social Security surpluses are invested in U.S.Treasury bonds, which are projected to providerates of return that exceed inflation by 2.3 per-cent.3 This is a low rate of return relative to theaverage above-inflation yield of about 7 percent onequities over the past several decades.4 Proponentsof private market investment of retirement taxesand/or of national defined contribution-stylesavings reason that investing in the equitiesmarket will provide better rates of return andthereby lower program costs. Some also believeswitching retirement tax investment into theprivate sector will increase net savings and providecapital to spur economic growth. Advocates ofallowing or mandating defined contribution-stylesavings within the national government retirementsystem similarly cite economic growth and in-creased savings as primary advantages. 5

■ Reform ProposalsAlthough many policymakers seem to agree thatprivate investment of retirement taxes in equitieswill be part of any reform, exactly how to investand to what extent has been a topic of hot debate.For example, the 13-member 1994–1996 SocialSecurity Advisory Council was assembled by theSecretary of Health and Human Services to proposea solution to the program’s financing needs. Thishighly factionalized council recently released amuch anticipated report delineating three reformpackages, each proposing different means ofinvesting varying proportions of retirement taxes inthe equities market. In fact, dissenting views areexpressed even among supporters of the samereform packages. Meanwhile, other reform propos-als have been forthcoming from both legislators andpublic policy organizations.

As a result of the multitude and complexity ofthe reform plans, it is difficult to stay abreast of thedebate. To provide a broad summary of thesereforms, table 1 shows a point-by-point comparisonof seven major Social Security reform packages.Keep in mind that some of the complexities andcaveats of these reforms have been pared down inorder to create a summary and that these plansmay continue to evolve as the debate progresses.

This table includes the best informationavailable at this time. A request for verification ofthe table’s content was sent to each reform sponsorand to several plan supporters. A number of plansponsors and supporters responded, and everyeffort was made to incorporate their revisions andsuggestions. The author welcomes additionalinformation and comments from readers.


114

■ Endnotes1 For the purposes of this article, the term retire-ment taxes only refers to revenue that is deductedfrom workers’ payroll earnings for the purpose offunding a national retirement program.

2 “Actuarial Status of the Social Security andMedicare Trust Funds,” Social Security Bulletin(Summer 1996): 61.

3 Henry Aaron “Behind the Privatization Rush,”Washington Post, July 22, 1996, p. C1.

4 Ibid.

5 For example, Jose Pinera of the Cato Instituteattributes Chile’s increased savings rate of 26percent of gross national product and its economicgrowth rate of 7 percent, double Chile’s historic rate,to Social Security privatization (Jose Pinera, “TheSuccess of Chile’s Privatized Social Security,” CatoPolicy Report (July/August 1995): 10. In addition,

Sylvester Schieber, co-sponsor of the PersonalSecurity Account Plan, states his plan would “turnour Social Security system into a major engine ofreal savings for workers so they can both securetheir own retirement income needs while alsomaking a contribution to the future growth of thenational economy” (Proposals for Retirement PolicyReform: Ensuring Our Workers’ Retirement Secu-rity, Testimony before the Senate Labor and HumanResources Committee, Aging Subcommittee, July16, 1996).

115

TY

PE O

F P

LA

N

IN

VESTM

EN

T P

LA

N

Private M

arket

In

vestm

en

t?

Who I

nvests?

In

dividual A

ccoun

ts

In

dividual

Participation

Where I

s I

ndividual

Accoun

t H

eld?

DB a

nd v

aria

ble

DC, w

ithde

cline

in D

B as

DC

com

pone

nt o

f new

sys

tem

grow

s.

Yes

Indi

vidu

als

Yes,

cal

led

pers

onal

retir

emen

t sav

ings

acco

unts

(PRS

As)

Optio

nal

Indi

vidu

als

plac

e DC

acco

unt i

n p

rivat

e m

arke

t.

DB a

nd D

C

Yes

Indi

vidu

als

and

fede

ral

gove

rnm

ent

Yes,

cal

led

pers

onal

inve

stm

ent p

lans

(PIP

s)

Optio

nal i

n S.

824

;m

anda

tory

in S

. 825

U.S.

Trea

sury

; lik

e Th

rift

Savi

ngs

Fund

, som

e PI

Psal

low

ed in

IRAb

-type

acco

unts

. Adm

inis

tere

d by

bank

s, c

redi

t uni

ons,

or

othe

r ins

titut

ions

sub

ject

to b

anki

ng la

ws.

DC w

ith a

redi

strib

utio

nal

com

pone

nt (s

ee b

elow

) for

low

ear

ning

wor

kers

and

a m

inim

um re

tirem

ent

inco

me

guar

ante

e; a

llOA

SIa

“el

der”

ben

efits

(i.e.

, ben

efits

for r

etire

dw

orke

rs, a

ged

spou

ses

and

wid

ow[e

r]s)

are

phas

ed o

ut.

Yes

Indi

vidu

als;

gov

ernm

ent i

sde

faul

t inv

esto

r for

thos

eun

able

or u

nwill

ing

eith

erto

man

age

thei

r ow

n fu

nds

or to

hire

a fi

nanc

ial

man

ager

to d

o so

.

Yes,

cal

led

pers

onal

thrif

tac

coun

ts (P

TAs)

Man

dato

ry, w

ith a

llow

ance

for e

xtra

vol

unta

ryco

ntrib

utio

ns

Indi

vidu

als

plac

e in

priv

ate

mar

ket

Phas

e ou

t one

-hal

f of

exis

ting

DB s

yste

m;

repl

ace

othe

r hal

f with

aDC

sys

tem

. Re

mai

ning

DB p

art o

f sys

tem

isre

stru

ctur

ed to

pro

vide

a fla

t ben

efit

for a

llpa

rtic

ipan

ts.

Yes

Indi

vidu

als

or th

eir

desi

gnat

ed b

roke

r or

finan

cial

inst

itutio

n

Yes,

cal

led

pers

onal

savi

ngs

acco

unts

(PSA

s)

Man

dato

ry fo

r wor

kers

unde

r age

55

in 1

998

In p

rivat

e in

stitu

tions

desi

gnat

ed b

y w

orke

rs

DB a

nd d

efin

edco

ntrib

utio

n (D

C)

Yes

Indi

vidu

als

Yes,

cal

led

indi

vidu

alac

coun

ts (I

As)

Man

dato

ry

Like

the

Fede

ral T

hrift

Plan

, acc

ount

s ar

e he

ld b

yth

e fe

dera

l gov

ernm

ent,

but a

re s

epar

ate

from

fede

ral g

over

nmen

tas

sets

and

are

not

use

d in

calc

ulat

ing

the

fede

ral

debt

.

Defin

ed b

enef

it (D

B)

Poss

ibly

Fede

ral g

over

nmen

t

No None

Not a

pplic

able

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

Appe

ndix

A: T

able

1A

Poin

t-

by-

Poin

t C

om

parison

of S

ocial S

ecurity R

eform

Plan

s

115

CED

Reform

Plan

DB a

nd D

C

Yes

Indi

vidu

als

Yes,

cal

led

pers

onal

retir

emen

t acc

ount

s(P

RAs)

Man

dato

ry

Priv

atel

y ow

ned

and

held


116

Yes;

inve

st p

art o

ftr

ust f

und

accu

mul

atio

nsth

at w

ould

oth

erw

ise

bein

vest

ed in

U.S

. Tre

asur

ybo

nds;

gra

dual

inve

stm

ent u

ntil

40%

of tr

ust f

und

accu

mul

a-tio

ns a

re in

vest

ed in

priv

ate

mar

ket s

ecur

ities

by 2

014.

Yes;

0.8

% o

n em

ploy

ers

and

empl

oyee

s(1

.6%

add

ition

al ta

xov

eral

l) be

ginn

ing

in20

45; n

on-M

edic

are

FICA

wou

ld th

en e

qual

14.0

% to

tal.

The

MB

plan

wou

ld in

crea

seco

ntrib

utio

ns b

y0.

15 p

erce

nt o

n th

eem

ploy

ee's

cove

red

wag

es, t

o be

mat

ched

by

the

empl

oyer

onl

y if

it di

dno

t cha

nge

the

com

puta

tion

perio

d.

Not a

pplic

able

Not a

pplic

able

No

Yes;

1.6

% o

n em

ploy

ees

in 1

998.

DB

part

rem

ains

fina

nced

by

12.4

% F

ICA;

tota

l FIC

A in

199

8 w

ould

be

14.0

%

To s

pous

e as

if jo

int

annu

ity

Empl

oyee

con

trib

utio

nsar

e af

ter t

ax, w

hile

dist

ribut

ions

are

free

of

fede

ral i

ncom

e ta

xes

or

tax

dedu

ctib

le w

hen

save

dan

d ta

xabl

e w

hen

bene

fits

are

paid

.

Yes;

inve

st 5

per

cent

age

poin

ts o

f the

non-

Med

icar

e FI

CAra

te (1

2.4%

), pa

id o

nly

by th

e em

ploy

ee(r

emai

ning

7.4

% u

sed

to fi

nanc

e th

e fla

t ben

efit,

as w

ell a

s su

rviv

ors,

spou

se, a

nd d

isab

ility

bene

fits)

.

Yes;

by

1.52

% o

nem

ploy

ers

and

empl

oyee

s fo

r 70

year

s,be

ginn

ing

in 1

998,

tohe

lp fi

nanc

e th

e tr

ansi

tion.

Yes

Indi

vidu

als’

con

trib

utio

nsto

PSA

s ar

e af

ter t

ax, s

odi

strib

utio

ns a

re ta

x fr

ee.

Yes;

any

ann

ual O

ASI

oper

atin

g su

rplu

s is

cred

ited,

pro

rata

, to

wor

kers

’ PIA

s.

No, c

urre

nt la

w F

ICA

rem

ains

initi

ally

unch

ange

d; o

nce

OASI

oper

atin

g su

rplu

ses

exce

ed 3

% o

f pay

roll,

FICA

taxe

s ar

e re

duce

d.Ho

wev

er, w

orke

rs a

rere

quire

d to

con

trib

ute

5%of

thei

r ear

ning

s to

PIA

s(n

o em

ploy

er c

ontr

ibu-

tions

); co

ntrib

utio

ns o

flo

w e

arne

rs a

re m

atch

edby

the

gove

rnm

ent,

with

the

subs

idy

phas

ed o

ut a

tth

ree

times

the

low

wag

eth

resh

old.

Yes

PTA

cont

ribut

ions

and

capi

tal a

ccum

ulat

ion

are

exem

pt fr

om p

erso

nal

inco

me

taxa

tion;

dist

ribut

ions

are

fully

taxa

ble.

In S

. 825

, yes

; inv

est

25%

of a

nnua

l tru

st fu

nd a

ccum

ulat

ions

and

2% o

f em

ploy

ee’s

shar

e of

non

surp

lus,

non-

Med

icar

e FI

CA; i

nS.

824

, gov

ernm

ent

does

not

inve

st tr

ust

fund

acc

umul

atio

ns, b

ut2%

of e

mpl

oyee

’s sh

are

of n

on-M

edic

are

FICA

ispu

t int

o PI

Ps o

n an

optio

nal b

asis

.

No Yes

Like

an

IRA,

PIP

cont

ribut

ions

and

accu

mul

atio

ns a

re ta

xfr

ee u

ntil

dist

ribut

ion.

Yes,

inve

st tr

ust f

und

accu

mul

atio

ns o

ver 1

year

’s w

orth

of b

enef

its;

spec

ific

% a

vaila

ble

for

inve

stm

ent i

s ba

sed

onho

w m

uch

peop

le ta

ke in

Soci

al S

ecur

ity D

Bbe

nefit

s, a

ffect

ing

how

muc

h su

rplu

s th

ere

isun

der 1

2.4%

FIC

A; h

owm

uch

peop

le ta

ke fr

om D

Bpa

rt d

epen

ds o

nac

cum

ulat

ion

rate

s in

PRSA

s; p

erce

ntag

e of

surp

lus

avai

labl

e fo

rin

vest

men

t is

expe

cted

tobe

2.3

% fo

r 199

8.

No Yes

Cont

ribut

ions

are

not

taxe

d; d

istr

ibut

ions

are

taxe

d sa

me

as S

ocia

lSe

curit

y in

com

e ru

les.

No

Yes,

by

1.5

perc

enta

gepo

ints

for b

oth

empl

oyer

san

d em

ploy

ees.

How

ever

,go

vern

men

t rec

eive

s no

addi

tiona

l rev

enue

s.

Yes

Like

a 4

01(k

)—co

ntrib

utio

ns a

nd e

arni

ngs

are

tax

free

unt

ildi

strib

utio

n

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

CED

Reform

Plan

Use F

ICA

Reven

ue

for I

nvestm

en

t?

IN

CR

EA

SE P

AY

RO

LL

CO

NTR

IB

UTIO

NS?

PR

OP

ER

TY

OF

BEN

EFICIA

RY

’S

ESTA

TE U

PO

N

DEA

TH

?

TA

X T

REA

TM

EN

T O

F

IN

DIV

ID

UA

L

ACCO

UN

T B

EN

EFITS

117

Sim

ilar t

o cu

rren

t IRA

optio

ns, a

sset

s ar

ein

vest

ed a

t wor

kers

’di

scre

tion

in U

.S.

secu

ritie

s, in

sura

nce

cont

ract

s, c

ertif

icat

es o

fde

posi

t, co

mm

on s

tock

s,an

d/or

oth

er in

stru

men

tsor

obl

igat

ions

sel

ecte

d by

qual

ified

pro

fess

iona

las

set m

anag

ers.

?

Unde

cide

d, b

utpr

obab

ly n

ot.

No

$2.3

3 tr

illio

n by

201

5

IN

VESTM

EN

T

RU

LES

FO

RM

OF

DISTR

IB

UTIO

N

RU

LES

PR

ER

ETIR

EM

EN

T

LO

AN

S?

ACCR

UA

L O

F

AD

DITIO

NA

L

FED

ER

AL D

EB

T

REQ

UIR

ED

?

AM

OU

NTS

EX

PECTED

TO

BE

ALLO

CA

TED

TO

IN

DIV

ID

UA

L

ACCO

UN

TS

If p

rivat

e in

vest

men

t, in

corp

orat

e bo

nds

and

equi

ty in

dex

fund

s.

Not a

pplic

able

Not a

pplic

able

No

Not a

pplic

able

Base

d on

Fed

eral

Thrif

tSa

ving

s Pl

an—

choo

sefr

om a

list

of i

ndex

fund

s ad

min

iste

red

byth

e So

cial

Sec

urity

Adm

inis

trat

ion

(SSA

).

Man

dato

ry m

inim

umin

dexe

d an

nuity

upo

nre

tirem

ent,

(man

dato

ryjo

int i

f mar

ried

unle

sssp

ouse

sig

ns a

wai

ver)

issu

ed u

nder

SSA

rule

san

d un

derw

ritte

n by

the

gove

rnm

ent.

No No

Appr

oxim

atel

y $

50 b

illio

nan

nual

ly/

$1.2

trill

ion

by th

e en

d of

201

5 (1

996

dolla

rs)

Asse

ts a

re in

vest

ed a

tth

e di

scre

tion

of th

ew

orke

r bu

t sub

ject

tore

gula

tions

rest

rictin

gth

em to

retir

emen

tpu

rpos

es a

nd to

inve

stm

ents

in w

idel

yhe

ld fi

nanc

ial

inst

rum

ents

.

Optio

nal a

nnui

tizat

ion

with

with

draw

al li

mits

base

d on

rem

aini

ng li

feex

pect

ancy

(not

fina

l)

No

Yes,

in a

dditi

on to

the

1.52

% ta

x in

crea

se,

“lib

erty

bon

ds”

wou

ld b

eis

sued

by

the

fede

ral

gove

rnm

ent f

rom

200

2th

roug

h 20

69 to

fina

nce

the

tran

sitio

n to

the

new

syst

em.

Libe

rty

bond

sw

ould

equ

al a

bout

$65

0bi

llion

by

2030

and

be

repa

id w

ith th

e pr

ocee

dsof

the

tran

sitio

n ta

x of

1.52

%. L

iber

ty b

onds

wou

ld b

e eq

ual t

o 1.

97%

of ta

xabl

e pa

yrol

l ann

ually

.

Appr

oxim

atel

y$3

.3 tr

illio

n by

the

end

of 2

015

(199

6 do

llars

)

Fed

eral

ove

rsig

ht b

oard

dete

rmin

es %

of a

sset

sre

quire

d in

risk

-free

deb

tan

d es

tabl

ishe

sdi

vers

ifica

tion

regu

latio

ns;

indi

vidu

als

sele

ct a

cert

ified

fina

ncia

l man

ager

of th

eir c

hoic

e, e

xcep

tth

at a

fede

ral b

oard

inve

sts

for w

orke

rsun

able

or u

nwill

ing

todo

so

them

selv

es.

Man

dato

ry a

nnui

tizat

ion

of m

inim

um a

nnui

ty;

full

free

dom

with

rem

aind

er o

f fun

ds

No

Yes,

a s

light

initi

alin

crea

se in

the

defic

itw

ould

pea

k at

$45

bill

ion

(or 0

.4%

of g

ross

dom

estic

prod

uct [

GDP]

) in

2006

;by

201

4, th

is d

efic

it tu

rns

into

a s

urpl

us.

Net a

nnua

l sav

ings

in P

TAs

rises

from

1.7

% o

f GDP

in 2

000

to 3

.4%

of G

DPby

203

0

Base

d on

Fed

eral

Thrif

tSa

ving

s Pl

an; o

ptio

n of

two

wid

e in

vest

men

tca

tego

ries:

(1) p

erso

nal

inve

stm

ent p

lan

with

low,

mod

erat

e, a

nd h

igh

risk

optio

ns, a

nd (2

) IRA

-like

acco

unts

with

inve

stm

ent

optio

ns e

xact

ly lik

ecu

rren

t IRA

opt

ions

.

Assu

med

man

dato

ryan

nuiti

zatio

n be

caus

e of

mod

elin

g af

ter F

eder

alTh

rift P

lan,

but

no

expl

icit

info

rmat

ion

foun

d

No No

In S

. 825

, 2%

of n

on-

Med

icar

e FI

CA; S

. 824

,op

tiona

l, so

allo

catio

nde

pend

s on

beh

avio

ral

resp

onse

to P

IP o

ptio

n.

Asse

ts a

re in

vest

ed in

alim

ited

num

ber o

f bro

ad-

base

d fu

nds

that

may

inve

st in

priv

ate-

sect

orfin

anci

al s

ecur

ities

,go

vern

men

t deb

t, et

c.

Fund

s m

ust b

e w

ithdr

awn

grad

ually

ove

r life

afte

rre

tirem

ent o

r ann

uitiz

ed

No

Fede

ral d

ebt r

educ

ed in

the

futu

re b

y m

akin

g DB

plan

sol

vant

.

Mor

e th

an $

100

billi

onan

nual

ly, a

ccum

ulat

ing

to$2

.5 tr

illio

n by

201

5(1

996

dolla

rs)

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

CED

Reform

Plan


118

Rais

es n

et n

atio

nal

savi

ngs

by 2

.6%

of G

DP b

y20

10 a

nd b

y 5%

of G

DP b

y20

30.

Roug

hly

one-

half

of P

TAas

sets

; 0.8

5% o

f GDP

annu

ally

in 2

000

risin

gto

1.7

% a

nnua

llyby

203

0.

Tota

l exp

ecte

d va

lue

ofpr

ivat

ely

inve

sted

pla

nas

sets

by

2015

isap

prox

imat

ely

35%

of G

DP,

risin

g to

55%

of G

DPby

203

0.

OASI

“el

der b

enef

its”

are

grad

ually

pha

sed

out b

y re

duci

ng n

ewly

earn

ed w

age

cred

itsus

ed in

cal

cula

ting

futu

re b

enef

its.

No

?

Appr

oxim

atel

y$2

0 to

$25

bill

ion

on a

vera

ge

$600

bill

ion

by 2

015

Not a

pplic

able

No

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

Not a

pplic

able

Appr

oxim

atel

y$2

5 bi

llion

on a

vera

ge

$1.1

trill

ion

by 2

015

Not a

pplic

able

Yes,

at t

he ti

me

Med

icar

eis

refin

ance

d; n

ote

that

redi

rect

ion

wou

ld a

mou

ntto

a re

duct

ion

in H

I fun

dseq

uiva

lent

to 0

.31%

of

taxa

ble

payr

oll.

?

Appr

oxim

atel

y$7

5 bi

llion

on a

vera

ge

$2.1

trill

ion

by 2

015

No “

offs

ettin

g”re

duct

ions

, but

not

e th

atw

orke

rs a

ged

25 th

roug

hag

e 54

in 1

998

wou

ldre

ceiv

e a

bene

fit th

at is

a co

mbi

natio

n of

the

exis

ting

curr

ent-l

awde

fined

ben

efit

and

new

flat b

enef

it, w

itham

ount

s pr

orat

ed to

refle

ct ti

me

cove

red

byol

d an

d ne

w s

yste

ms.

Yes;

imm

edia

tely;

not

eth

at re

dire

ctio

n w

ould

amou

nt to

a re

duct

ion

inHI

fund

s eq

uiva

lent

to0.

31%

of t

axab

le p

ayro

ll.

No d

ata

foun

d on

S.8

25;

S.82

4 is

act

uaria

llyne

utra

l.

? ?

In S

. 824

, PIP

accu

mul

atio

ns a

re o

ffset

with

cor

resp

ondi

ngre

duct

ions

in b

end

poin

ts s

o th

at th

e pl

anis

act

uaril

y ne

utra

l.

Yes;

not

e th

at re

dire

ctio

nw

ould

am

ount

to a

redu

ctio

n in

HI f

unds

equi

vale

nt to

0.3

1% o

fta

xabl

e pa

yrol

l.

?

$78.

1 bi

llion

in 1

998;

aver

age

per y

ear 1

998

thro

ugh

2015

is$1

29.5

bill

ion.

Maj

ority

of P

RSA

bala

nces

exp

ecte

d to

go in

to p

rivat

e ca

pita

lm

arke

t (se

e ab

ove)

.

For P

RSA

part

icip

ants

,pr

esen

t val

ue o

fex

pect

ed li

fetim

e So

cial

Secu

rity

bene

fits

isre

duce

d by

the

pres

ent

valu

e in

the

PRSA

,di

scou

nted

at a

rate

of 4

.3%

Yes;

not

e th

at re

dire

ctio

nw

ould

am

ount

to a

redu

ctio

n in

HI f

unds

equi

vale

nt to

0.3

1% o

fta

xabl

e pa

yrol

l.

EX

PECTED

IN

CR

EA

SE

IN

NA

TIO

NA

L

SA

VIN

GS?

HO

W M

UCH

NET

AD

DITIO

NA

L

REV

EN

UE I

S

EX

PECTED

TO

FLO

W

IN

TO

STO

CK

MA

RK

ET

AN

NU

ALLY

IN

19

96

DO

LLA

RS?

C

HO

W M

UCH

NET

AD

DITIO

NA

L

REV

EN

UE I

S

EX

PECTED

TO

FLO

W

IN

TO

PR

IV

ATE

CA

PITA

L M

AR

KET B

Y

20

15

IN

19

96

DO

LLA

RS?

OFFSETTIN

G

RED

UCTIO

N I

N

SO

CIA

L S

ECU

RITY

BECA

USE O

F P

RIV

ATE

IN

VESTM

EN

T

MISCELLA

NEO

US

REFO

RM

S

Redirect S

ocial

Security T

ax

Reven

ues f

rom

Medicaid P

art A

Hospital I

nsuran

ce

(H

I) t

o t

he S

ocial

Security T

rust

Fun

d?

d

Larg

e in

crea

sean

ticip

ated

; pre

cise

amou

nt is

unk

now

able

.

Appr

oxim

atel

y $5

0 bi

llion

on a

vera

ge

$1.1

trill

ion

by 2

015

None No

CED

Reform

Plan

119

Cover N

ew

State a

nd

Local G

overn

men

t

Em

ployees?

BEN

EFIT

RED

UCTIO

NS F

OR

FU

TU

RE R

ETIR

EES

Raise M

in

im

um

TA

XES o

n B

en

efits?

Phase O

ut I

ncom

e

Thresholds?

In

crease N

orm

al

Retirem

en

t A

ge

(N

RA

) F

aster T

han

Chan

ge M

an

dated b

y

the 1

98

3

Am

en

dm

en

ts (

age 6

7

by 2

02

7)

Yes

Yes,

fede

rally

tax

all

bene

fits

from

DB

port

ion

in e

xces

s of

con

trib

utio

nsm

ade

by w

orke

r. Thi

sw

ould

be

achi

eved

by

mak

ing

85%

of S

ocia

lSe

curit

y be

nefit

s ta

xabl

ein

com

e, n

ot b

y ca

lcul

atin

gfe

dera

l tax

es in

divi

dual

lyfo

r ben

efic

iarie

s.

Yes

Yes,

rais

e NR

A by

2 m

onth

s ea

ch y

ear u

ntil

itre

ache

s ag

e 70

in 2

030.

Yes

Yes,

tax

as g

ener

alre

venu

e

No

Yes,

in 2

002,

NRA

rise

s b

y th

ree

mon

ths

per

year

unt

il ag

e 69

in20

00 to

201

8, in

dexa

tion

ther

eafte

r to

long

evity

.

Yes

For a

ll ag

ed 5

5 or

ove

r,on

e-ha

lf of

ben

efits

wou

ldbe

taxa

ble;

all T

ier 1

(db)

bene

fits

wou

ld b

e ta

xabl

e;el

imin

ate

the

AGI (

adju

sted

gros

s in

com

e) th

resh

old;

on D

C pa

rt, c

ontr

ibut

ions

are

afte

r tax

but

tax

free

upon

dis

trib

utio

n. Th

ein

com

es o

f low

-wag

eea

rner

s w

ould

stil

l be

prot

ecte

d, a

s ab

out 3

0pe

rcen

t of S

ocia

l Sec

urity

bene

ficia

ries

wou

ld n

ot p

ayan

y ta

xes.

Yes

Yes,

gra

dual

ly in

crea

seNR

A by

2 m

onth

s an

nual

lybe

ginn

ing

in 2

000

until

age

67 fo

r tho

se re

achi

ng a

ge62

by

2011

; ind

ex to

long

evity

ther

eafte

r(p

roje

cted

to b

e ab

out 1

mo.

eve

ry 2

yrs.

); th

is ra

teof

incr

ease

to b

e re

view

edev

ery

10 yr

s. b

y Co

ngre

ss.

Yes

Yes,

tax

unde

r gen

eral

inco

me

tax

prin

cipl

es;

taxe

s ca

lcul

ated

indi

vidu

ally

; ave

rage

taxa

ble

inco

me

wou

ld b

e92

% .

All S

ocia

l Sec

urity

bene

fits

in e

xces

s of

alre

ady

taxe

d em

ploy

eeco

ntrib

utio

ns w

ould

be

incl

uded

in fe

dera

l tax

able

inco

me

and

the

proc

eeds

depo

site

d to

the

OASD

Itr

ust f

unds

. The

inco

me

oflo

w-w

age

earn

ers

wou

ldst

ill b

e pr

otec

ted,

as

abou

t30

per

cent

of S

ocia

lSe

curit

y be

nefic

iarie

sw

ould

not

pay

any

taxe

s.

Yes

No

Yes

Yes,

tax

unde

r gen

eral

inco

me

tax

prin

cipl

es;

taxe

s ca

lcul

ated

indi

vidu

ally

; ave

rage

taxa

ble

inco

me

wou

ld b

e92

% o

n DB

par

t; on

DC

part

, con

trib

utio

ns a

reaf

ter

tax

but t

ax fr

eeup

on d

istr

ibut

ion.

All

Soci

al S

ecur

ity b

enef

its in

exce

ss o

f alre

ady

taxe

dem

ploy

ee c

ontr

ibut

ions

wou

ld b

e in

clud

ed in

fede

ral t

axab

le in

com

ean

d th

e pr

ocee

dsde

posi

ted

to th

e OA

SDI

trus

t fun

ds. T

he in

com

esof

low

-wag

e ea

rner

sw

ould

stil

l be

prot

ecte

d,as

abo

ut 3

0 pe

rcen

t of

Soci

al S

ecur

itybe

nefic

iarie

s w

ould

not

pay

any

taxe

s.

Yes

Yes,

gra

dual

ly in

crea

seNR

A by

2 m

onth

san

nual

ly b

egin

ning

in20

00 u

ntil

age

67 fo

rth

ose

reac

hing

age

62

by20

11; i

ndex

to lo

ngev

ityth

erea

fter(

proj

ecte

d to

be a

bout

1 m

o. e

very

2 yr

s.);

full

bene

fit N

RAeq

uals

age

69

by 2

059.

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

CED

Reform

Plan

Poss

ibly

85%

of a

ll OA

SDI b

enef

itssu

bjec

t to

pers

onal

inco

me

taxe

s.

Yes

No

Yes

No No

Yes,

to a

ge 7

0 by

202

9(S

. 825

); in

dexe

d to

long

evity

ther

eafte

r.


120

Yes,

incr

emen

tal

incr

ease

s be

ginn

ing

in20

02 u

ntil

ERA

is 6

5 by

2017

; ind

ex a

fter 2

030.

Yes,

by

0.5

% p

oint

s; li

mit

full

COLA

s to

ben

efic

iarie

sbe

low

30t

h pe

rcen

tile;

thos

e ab

ove

30th

perc

entil

e ge

t onl

y th

eCO

LA a

mou

nt d

ue th

ose

atth

e 30

th p

erce

ntile

.

Yes,

cha

nge

bend

poi

nts

to b

e m

ore

prog

ress

ive

(two

diffe

rent

way

s to

do th

is in

S.8

24 a

ndS.

825)

Yes,

cha

nge

from

am

axim

um o

f 50%

of

retir

ed s

pous

es’ f

ull

bene

fit to

33%

max

imum

.

No

Yes,

by

0.5

perc

enta

gepo

ints

, the

am

ount

anno

unce

d by

the

Bure

auof

Lab

or S

tatis

tics

inM

arch

199

6.

Annu

al w

age

cred

its fo

rOA

SI “

elde

r” b

enef

itsar

e re

duce

d by

5%

per

year

so

that

no

bene

fits

are

bein

g ea

rned

afte

r20

yea

rs; i

n ef

fect

, thi

sam

ount

s to

a p

ro ra

taco

hort

redu

ctio

n in

bene

fits

whe

re a

mou

ntof

redu

ctio

n de

pend

s on

year

ben

efic

iarie

s w

ere

born

(or t

urn

age

62).

New

spo

usal

ben

efits

are

phas

ed o

ut o

ver

10 y

ears

.

Yes,

gra

dual

ly in

crea

se to

age

65 in

the

year

203

5 fo

rDB

par

t (no

t ind

exed

) but

DC p

art c

ould

be

draw

non

as

early

as

age

62.

No, b

ut a

ssum

es th

e an

nual

CPI

is re

duce

d fr

omTr

uste

es’ a

ssum

ptio

ns b

y0.

21 p

erce

ntag

e po

ints

,th

e am

ount

ann

ounc

ed b

yth

e Bu

reau

of L

abor

Stat

istic

s in

Mar

ch 1

996.

Yes,

for D

isab

ility

Insu

ranc

e (s

ee b

elow

);al

so, t

he “

Soci

alSe

curit

y” p

ortio

n (th

efla

t ben

efit)

wou

ld n

owbe

bas

ed o

n ye

ars

ofco

vera

ge, n

ot a

vera

geea

rnin

gs; r

etire

men

tea

rnin

gs te

st is

elim

inat

ed b

y 20

02.

No, s

pous

es w

ould

rec

eive

the

high

er o

f th

eir o

wn

retir

emen

t ben

efit,

one

-hal

fof

thei

r ret

ired

spou

se’s

bene

fit, o

r one

-hal

f of t

hefla

t ben

efit

(whe

n fu

llyph

ased

in);

spou

ses

wou

ldre

ceiv

e pr

ocee

ds fr

om th

eir

own

PSAs

.

No

No, b

ut a

ssum

es th

ean

nual

CPI

is re

duce

dfr

om Tr

uste

es’ a

ssum

p-tio

ns b

y 0.

21 p

erce

ntag

epo

ints

, the

am

ount

anno

unce

d by

the

Bure

auof

Lab

or S

tatis

tics

inM

arch

199

6.

Yes,

bes

t 35

year

s to

38 y

ears

;e a

nd g

radu

ally

scal

e ba

ck b

enef

its fo

rea

rner

s in

the

seco

ndan

d th

ird ta

x br

acke

ts b

ych

angi

ng fo

rmul

a (1

) so

that

ben

efits

can

be

finan

ced

with

cur

rent

12.4

% p

ayro

ll ta

x an

d (2

) so

as to

pre

serv

ecu

rren

t ben

efit

leve

l in

real

dol

lars

.f

Yes,

to 3

3% o

f PIA

(als

o se

e su

rvivo

r’sbe

nefit

s).

No

No, b

ut a

ssum

es th

ean

nual

CPI

is

redu

ced

from

Trus

tees

’as

sum

ptio

ns b

y 0.

21pe

rcen

tage

poi

nts,

the

amou

nt a

nnou

nced

by

the

Bure

au o

f Lab

or S

tatis

tics

in M

arch

199

6.

Yes,

bes

t 35

year

s to

bes

t 38e

yea

rs (e

qual

s,on

ave

rage

, a 3

%re

duct

ion

in b

enef

itsby

199

9). T

he M

B pl

anw

ould

eith

er c

hang

e th

eco

mpu

tatio

n pe

riod

orin

crea

se c

ontr

ibut

ions

by

0.15

per

cent

on

the

empl

oyee

's co

vere

dw

ages

, to

be m

atch

ed b

yth

e em

ploy

er.

No

In

crease E

arly

Retirem

en

t A

ge

(ER

A)?

Reduce C

ost-

of-

Livin

g

Adjustm

en

t (

CO

LA

)

by R

educin

g

Con

sum

er P

rice

In

dex (

CP

I)?

Chan

ge i

n B

en

efit

Calculation

s?

Reduce S

pousal

Ben

efits?

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

CED

Reform

Plan

Yes,

gra

dual

ly to

age

65

from

200

0 th

roug

h 20

11;

trac

ks N

RA to

long

evity

afte

r 201

5, b

ut fo

ur y

ears

beh

ind.

No

Best

39

year

s in

stea

dof

bes

t 35

year

se ; a

lso,

cha

nge

bend

poin

ts to

be

mor

epr

ogre

ssiv

e.

Yes,

cha

nge

from

am

axim

um o

f 50%

of

retir

ed s

pous

es’ f

ull

bene

fit to

33%

max

imum

.

No

No, b

ut fa

vors

impr

ove-

men

t of a

ccur

acy

of C

PIm

easu

rem

ent b

eyon

dal

read

y im

plem

ente

d by

BLS

in 1

996

0.21

per

cent

corr

ectio

n.

Yes,

slo

w th

e in

crea

se o

fth

e PI

A as

link

ed w

ith re

alw

ages

for a

ll bu

t low

erin

com

e w

orke

rs. I

nad

ditio

n, c

alcu

late

from

best

40

year

s ra

ther

than

best

35

year

s.e

Yes,

gra

dual

ly fr

om 1

/2 o

fw

orke

rs P

IA to

1/3

.

121

Effect o

n D

isability

In

suran

ce

(D

I) a

nd

Survivors

In

suran

ce (

SI)

MIN

IM

UM

GU

AR

AN

TEED

BEN

EFIT T

O A

LL W

HO

PA

RTICIP

ATE?

BEN

EFIT

RED

UCTIO

NS A

ND

/OR

BEN

EFIT T

AX

IN

CR

EA

SES F

OR

CU

RR

EN

T

BEN

EFICIA

RIES

No c

hang

e in

DI;

sur

vivi

ngsp

ouse

s in

herit

PIP

accu

mul

atio

ns.

No c

hang

e fr

omcu

rren

t law

.h

COLA

cha

nges

wou

ld a

pply

to c

urre

nt b

enef

icia

ries,

but n

o on

e ov

er a

ge 5

1w

ould

be

affe

cted

by

age

chan

ge.

Surv

ivor

s w

ould

inhe

ritPR

SA a

mou

nts;

prim

ary

insu

ranc

e a

mou

nt (P

IA)g

wou

ld n

ot b

e re

duce

d fo

rpe

ople

with

PRS

As w

hoar

e di

sabl

ed a

nd u

nder

age

60.

No c

hang

e fr

omcu

rren

t law

.h

Uppe

r inc

ome

bene

ficia

-rie

s w

ho h

ave

alre

ady

gotte

n ou

t of t

he s

yste

mw

hat t

hey

and

thei

rem

ploy

ers

cont

ribut

edha

ve b

enef

its c

ut; n

one

over

age

58

wou

ld b

eaf

fect

ed b

y NR

A/ER

Ain

crea

ses;

no

one

rece

ivin

g sp

ousa

l ben

efits

befo

re 2

000

wou

ld b

eaf

fect

ed.

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

CED

Reform

Plan

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

Grad

ually

redu

ce D

Ipr

imar

y be

nefit

pay

able

as

NRA

incr

ease

s to

ens

ure

DI b

enef

its d

o no

t exc

eed

bene

fits

avai

labl

e th

roug

hea

rly re

tirem

ent;

resu

ltsin

a 3

0% re

duct

ion

in D

Ibe

nefit

s by

203

8 (h

owev

er,

over

all D

I red

uctio

nsw

ould

nev

er fa

ll be

low

70%

of c

urre

nt la

w);

incr

ease

age

d su

rvivo

rsbe

nefit

s (=

75%

com

bine

dor

100

% d

ecea

sed

wor

ker’s

ben

efit)

; DI a

ndSI

stil

l adm

inis

tere

d by

SSA;

no

chan

ge in

you

ngsu

rvivo

rs’ b

enef

it.

Yes,

for w

orke

rspa

rtic

ipat

ing

for 3

5 ye

ars,

equi

vale

nt to

$41

0 pe

rm

onth

in b

enef

its (1

996

dolla

rs) f

rom

Tier

I (D

Bpa

rt),

inde

xed

by g

row

thof

ave

rage

wag

es; a

smal

ler m

inim

um b

enef

itfo

r wor

kers

par

ticip

atin

gfe

wer

yea

rs; p

eopl

e w

ithon

ly 1

0 ye

ars

part

icip

atio

nge

t one

-hal

f of f

ull-c

aree

rw

orke

r’s b

enef

it.

Gran

dfat

her i

n pe

rson

sag

e 55

+ in

199

8 bu

tw

ith n

ew N

RA, E

RA,

and

tax

rule

s.

Surv

ivor’s

ben

efits

wou

ldbe

hig

her o

f 75%

of

com

bine

d co

uple

s be

nefit

or 1

00%

of d

ecea

sed

wor

ker’s

ben

efit.

No c

hang

e fr

omcu

rren

t law

.h

No c

hang

e in

gro

ssbe

nefit

s, b

ut n

et b

enef

itsar

e re

duce

d by

impl

emen

ting

new

tax

rule

s (s

ee a

bove

).

No c

hang

e fr

omcu

rren

t law

.

No c

hang

e fr

omcu

rren

t law

.h

No c

hang

e in

gro

ssbe

nefit

s, b

ut n

et b

enef

itsar

e re

duce

d by

impl

emen

t-in

g ne

w ta

x ru

les

(see

abov

e).

DI is

una

ffect

ed, e

xcep

tth

at b

enef

its a

re fu

llyta

xed;

SI i

s ph

ased

out

exce

pt fo

r ben

efits

toch

ildre

n, w

idow

edm

othe

rs a

nd fa

ther

s,an

d no

nage

d di

sabl

edw

idow

(er)

s.

Yes,

all

Amer

ican

s ag

ed62

and

ove

r are

guar

ante

ed a

tota

lho

useh

old

inco

me

equa

lto

100

per

cent

of t

hepo

vert

y lin

e.

One

year

COL

A fr

eeze

,w

ith C

OLAs

ther

eafte

r set

at th

e CP

I min

us 0

.5pe

rcen

tage

poi

nts;

85%

of

all O

ASDI

ben

efits

are

subj

ect t

o pe

rson

alin

com

e ta

xes.

Stat

emen

t doe

s no

tad

dres

s DI

.

No c

hang

e fr

om c

urre

ntla

w.

No c

hang

e in

gro

ssbe

nefit

s, b

ut n

et b

enef

itsar

e re

duce

d by

impl

emen

ting

new

tax

rule

s (s

ee a

bove

).


122

Person

al I

nvestm

en

t

Plan

Act o

f 1

99

5

(S. 8

24

) a

nd

Stren

gthen

in

g S

ocial

Security A

ct o

f 1

99

5

(S.8

25

) (

Kerrey-

Sim

pson

)

The S

ocial S

ecurity

Solven

cy A

ct

(R

ep. N

ick S

mith)

Nation

al T

axpayers’

Un

ion

Foun

dation

(N

TU

F) N

ation

al

Thrift P

lan

Person

al S

ecurity

Accoun

t P

lan

(W

eaver-

Schieber)

In

dividual A

ccoun

ts

Reform

Plan

(G

ram

lich)

Main

ten

an

ce o

f

Ben

efits R

eform

Plan

(B

all)

Im

plem

en

tation

Date

DO

PLA

N P

RO

PO

NEN

TS

EX

PECT P

LA

N T

O

RESO

LV

E 7

5 Y

EA

R

PR

OJECTED

SH

OR

TFA

LL?

CED

Reform

Plan

a Old-

Age

and

Surv

ivors

Insu

ranc

e.b In

divid

ual r

etire

men

t acc

ount

.c Al

l num

erica

l dat

a in

this

row

, exc

ept f

or th

e Na

tiona

l Tax

paye

rs’ U

nion

Fou

ndat

ion

plan

, wer

e pr

ovid

ed b

y St

eve

Goss

, Offi

ce o

f the

Act

uary

, Soc

ial S

ecur

ity A

dmin

istr

atio

n.d At

pre

sent

, a p

ortio

n of

Old

Age

Sur

vivor

s an

d Di

sabi

lity

tax

reve

nue

is a

lloca

ted

to fi

nanc

e M

edica

re’s

Par

t A (H

ospi

tal I

nsur

ance

) fun

d.e Cu

rren

tly, S

ocia

l Sec

urity

ben

efits

are

com

pute

d in

a fo

rmul

a us

ing

the

aver

age

of a

wor

ker’s

hig

hest

pai

d 35

yea

rs o

f wor

k. T

his

chan

ge w

ould

inclu

de th

ree

addi

tiona

l yea

rs in

to th

is fo

rmul

a. F

or m

ost w

orke

rs, c

hanc

es a

re th

e ad

ditio

n of

thes

eth

ree

year

s w

ill lo

wer

thei

r ave

rage

and

ther

eby

redu

ce th

eir S

ocia

l Sec

urity

ben

efits

, on

aver

age,

by

thre

e pe

rcen

t.f Th

e 32

per

cent

and

15

perc

ent c

onve

rsio

n fa

ctor

s in

the

pres

ent b

enef

it sc

hedu

le w

ould

be

grad

ually

low

ered

ove

r tim

e to

22.

4 pe

rcen

t and

10.

5 pe

rcen

t, re

spec

tivel

y.g Th

e pr

imar

y in

sura

nce

amou

nt is

cal

cula

ted

to d

eter

min

e OA

SDI b

enef

its. T

his

mea

ns n

o be

nefit

form

ula

chan

ges

wou

ld a

pply

for p

eopl

e w

ith p

erso

nal r

etire

men

t sav

ings

acc

ount

s w

ho a

re u

nder

age

60

and

disa

bled

.h Cu

rren

t law

doe

s no

t pro

vide

a sp

ecifi

c do

llar b

enef

it m

inim

um to

all

bene

ficia

ries,

but

it d

oes

prov

ide

a m

inim

um b

enef

it in

that

the

bene

fit fo

rmul

a is

com

pute

d us

ing

a sp

ecia

l min

imum

PIA

for p

erso

ns w

ith m

any

year

s of

low

wag

es.

i The

CED

plan

bui

lds

in a

mar

gin

for p

roje

ctio

n er

ror b

y id

entif

ying

bene

fit c

uts

suf

ficie

nt to

elim

inat

e th

e 13

3 pe

rcen

t of t

he p

rese

ntly

proj

ecte

d 75

-yea

r act

uaria

l def

icit.

1998

, but

inve

stm

ent o

fSo

cial

Sec

urity

trus

t fun

dac

cum

ulat

ions

wou

ld n

otbe

gin

until

200

0.

Only

if tr

ust f

unds

are

inve

sted

.

1998

Yes

1998

Yes

1998

Yes

1998

Yes

1998

Yes

Enac

ted

a.s.

a.p.

and

phas

ed in

gra

dual

ly.

Yes,

sol

ves

mor

e th

an10

0% o

f pro

ject

ed s

hort

falls

.

123

■ IntroductionReform of the Social Security system is now thesubject of serious discussions in various circles.According to the latest Social Security Trustees’report, cash flow1 in the program will turn negativein the year 2015, and the existing trust fund will bedepleted by 2029, under intermediate assumptions.Under the pessimistic assumptions, these dates are2000 and 2016, respectively; historically, thepessimistic assumptions have frequently proven tobe accurate. Current proposals for reform run thegamut from relatively minor adjustments to theexisting system to major restructuring. The SocialSecurity Advisory Council recently released threedifferent proposals for reform, and serious debate isnow beginning among policymakers, possibly withthe formation of some type of commission to makerecommendations.

■ A General Lack of ConfidenceAccording to the 1996 Retirement ConfidenceSurvey, co-organized by the Employee BenefitResearch Institute, the American Savings Educa-tion Council, and Mathew Greenwald and Associ-ates, both workers and retirees appear pessimisticregarding Social Security, but workers are muchmore so. Thirty-eight percent of retirees reportSocial Security as their most important income

Daring to Touch the Third Railby Paul Yakoboski

source, and an additional 26 percent consider it amajor source. By contrast, only 10 percent ofcurrent workers expect Social Security to be theirmost important source of retirement income, andonly 16 percent expect it to be a major source.Furthermore, 23 percent of workers do not expectSocial Security to be a source of income for them inretirement. Seventy-nine percent of workers arenot confident that the Social Security system willcontinue to provide benefits of value equal to thebenefits provided today, compared with 48 percentof retirees who feel the same way.

■ The Public Views ReformAccording to the 1996 Retirement ConfidenceSurvey, the public generally opposes reforms thatin some sense involve sacrifice, i.e., benefit cutsand/or tax increases. At the same time, the publictends to favor reforms that effectively trim benefitsfor higher income retirees. The public is alsoreceptive to more dramatic reforms, especiallythose that hold out the possibility of a “free lunch.”

Benefit Cuts and Tax Increases

The majority of both workers and retirees opposeany type of benefit cut. Majorities of both groupsstrongly oppose cutting future benefit payments forall future recipients (56 percent of workers and55 percent of retirees) (appendix B, chart 1). Anadditional 22 percent of workers and 14 percent ofretirees somewhat oppose such cuts.

Benefits can also be cut indirectly by raisingthe retirement age and/or by scaling back cost-of-living adjustments (COLAs), benefit increases thatoccur automatically with inflation. About60 percent of each group opposes reducing the levelof the automatic COLAs that occur with inflation.However, retirees are more strongly opposed toCOLA cutbacks than workers; 45 percent of retir

BAppendix

1 Cash flow is the difference between program income,excluding interest, and program outgo. Income exclud-ing interest consists of payroll-tax contributions,income from taxation of benefits, and miscellaneousreimbursements from the general fund of the U.S.Treasury. Outgo consists of benefit payments, adminis-trative expenses, net transfers from the Old-Age andSurvivors Insurance and Disability Insurance TrustFunds to the Railroad Retirement program under thefinancial-interchange provisions, and payments forvocational rehabilitation services for disabled benefi-ciaries.


124

Workers’ Attitudes Regarding

Social Security Reform

0 10 20 30 40 50 60 70 80 90 100Percentage

22% 42% 18% 13% 4%

28% 32% 19% 19% 2%

31% 32% 17% 18% 2%

28% 41% 13% 14% 4%

6% 22% 24% 45% 2%

4% 15% 22% 56% 3%

11% 23% 32% 31% 3%

9% 18% 24% 48% 1%

Retirees’ Attitudes Regarding

Social Security Reform

0 10 20 30 40 50 60 70 80 90 100Percentage

8% 32% 23% 26% 11%

37% 30% 9% 18% 5%

28% 28% 14% 20% 9%

16% 34% 13% 23% 13%

7% 24% 25% 34% 9%

4% 19% 14% 55% 9%

12% 18% 16% 45% 9%

13% 29% 20% 35% 4%

Partially “privatize” the system with individually directed accounts

Fully tax benefits for retirees with annual incomes over $50,000

Cut future benefits for retirees with annual incomes over $50,000

Invest some of the Trust Fund in the private-sector stock market

Increase the existing payroll tax on workers

Cut future benefits for all future recipients

Reduce automatic benefit increases that occur with inflation

Raise normal retirement age for full benefits to 70

Partially “privatize” the system with individually directed accounts

Fully tax benefits for retirees with annual incomes over $50,000

Cut future benefits for retirees with annual incomes over $50,000

Invest some of the Trust Fund in the private-sector stock market

Increase the existing payroll tax on workers

Cut future benefits for all future recipients

Reduce automatic benefit increases that occur with inflation

Raise normal retirement age for full benefits to 70

Appendix B: Chart 1

S

Strongly Favor

S

Somewhat Favor Somewhat Oppose

S

Strongly Oppose

D

Don’t Know

Source: Employee Benefit Research Institute/Mathew Greenwald and Associates/American Savings Education Council, 1996Retirement Confidence Survey.

125

ees strongly oppose such change, comparedwith31 percent of workers. Majorities of both groupsoppose raising the retirement age, but, not surpris-ingly, current workers are more strongly opposed.Seventy-two percent of workers and 55 percent ofretirees oppose raising the normal retirement agefor collecting full benefits to age 70.

In addition, majorities of both groups opposeincreasing the existing payroll tax on workers.Sixty-nine percent of workers oppose such reform,with 45 percent strongly opposed. Fifty-ninepercent of current retirees oppose such reform, withone-third (34 percent) strongly opposed.

Both groups favor cutting benefits for retireeswith higher incomes. Among retirees, 67 percentfavor fully taxing benefit payments of retirees withincomes over $50,000, and 56 percent favor cuttingfuture benefit payments for retirees with incomesover $50,000. Among workers, such proposals arefavored by 60 percent and 63 percent, respectively.Interestingly, even majorities of workers withannual incomes over $50,000 favor such proposals.

Trust Fund Equity Investment

Respondents were also asked about investing someof the trust fund in the stock market as opposed tokeeping it all invested in government securities.Sixty-nine percent of workers favored such change(with 28 percent strongly favoring). One-half ofretirees favored such reform. Over the long term,the stock market has historically had greaterreturns than government bonds. Apparentlyindividuals like the idea of the trust fund earning agreater rate of return than it currently does. Insome sense, they are expressing a preference forthe proverbial “free lunch,” but they may notunderstand the short-term volatility that comeswith equity investing and what this may mean forthe system.

Individual Accounts

Finally, respondents were asked their opinion of aproposal to deposit a fraction of their payroll taxesin an individual account over which they wouldexercise investment control. Income generated fromtheir account, combined with a guaranteed base

benefit, would then constitute their total SocialSecurity benefit. Benefits could be greater or lessthan those currently provided under the presentsystem.2 This is the one reform proposal on whichthe majority of workers and retirees disagreed.Sixty-four percent of workers favored such reform,with 22 percent strongly favoring it. Only 40percent of current retirees favored such reform.Obviously, such reform would not impact thosealready receiving benefits from the program. Theolder the worker, the less likely he or she was tosupport this proposal. Support did not vary notablywith worker income levels.

■ ImplicationsAmericans’ opinions regarding the Social Securitysystem are paradoxical. They are generally pessi-mistic about the current system and its ability tomaintain benefit levels into the future. At the sametime, they generally oppose any form of benefit cuts(except for higher income retirees) and/or taxincreases. What types of changes would they bewilling to accept? Apparently, investment of sometrust fund assets in private equity markets and thecreation of individual Social Security 401(k)-likeaccounts as part of the system. This is likelyindicative of two phenomena: first, a desire forrelatively painless solutions and, second, somedegree of distrust of the federal government in thisarea. Workers would be willing to have some of themoney in their name and under their control, asopposed to trusting in government promises thatare subject to change. This is the environmentwithin which elected officials must tackle thesituation.

2 The exact question wording was: “Currently, allSocial Security taxes in excess of those needed to paycurrent benefits are invested by the government ingovernment bonds. Some people have proposed thatindividuals be allowed to decide how some of themoney they pay in Social Security taxes is invested.Upon retirement, individuals would receive a reducedguaranteed Social Security payment, but they wouldalso receive income from the investments they chose.The total of these two sources of money could be higherthan current guaranteed benefits if the individual’sinvestments did well or lower if the individual’sinvestments did not do well. How do you feel about thisproposal?”


126

127

Public Attitudes on Social SecurityReformby Pamela Ostuw

■ IntroductionSurveys conducted by the Employee BenefitResearch Institute and the Gallup Organization,Inc. from February 1990 to March 1995 examinedpublic attitudes on Social Security.1 Time trendsand direct comparison among the surveys areproblematic due to changes in question wording orresponse options. However, the survey results doprovide some insights into attitudes toward SocialSecurity and how these attitudes have shifted overthe years.

■ The Current SystemIn the early 1990s, Americans were evenly split intheir beliefs about the likelihood that the SocialSecurity system will be able to pay benefits torespondents when they retire; in 1990 and 1991,49 percent believed they would receive benefits.However, in 1990, 92 percent of respondents did notbelieve that the Social Security benefits alonewould allow them to meet all of their financialneeds during retirement. In recent years, mostAmericans have become aware of the financingissues facing today’s Social Security system. By1995, 82 percent agreed or strongly agreed with thestatement that working Americans are beginning to

CAppendix

lose faith in whether Social Security benefits willbe available when they retire (appendix C, chart 1).

Is the current Social Security system a goodprogram for today’s younger workers? Only one-third of respondents in 1995 either agreed orstrongly agreed that it is, while nearly one-half(47 percent) disagreed or disagreed strongly.

■ Reform ProposalsSeveral of the reform proposals put forth todayadvocate contributions to individual retirementaccounts. In 1991, when asked if Social Securitytaxes, or a portion of these taxes, should go toindividual retirement accounts in the worker’s ownname, or if the system should remain as it is,61 percent thought the money should go to indi-vidual accounts, while 32 percent believed thesystem should stay as it is. The March 1995 surveyfound that 53 percent agreed or strongly agreedthat most people could make more money byinvesting their retirement funds in the privatesector than they could from Social Security. Thishas been a hot topic recently, with regard toindividuals’ ability to invest wisely and at anappropriate risk level. These concerns translateinto a concern for overall retirement incomeadequacy.

The idea of a voluntary Social Securitysystem has also arisen in the proposals for reform.In 1990 and 1991, among survey respondents, 45percent and 50 percent, respectively, were in favorof voluntary participation. While some advocates ofreform favor a voluntary program, this idea hasraised concern among others regarding the ad-equacy of individuals’ income in retirement.

When respondents were asked in the 1991survey if they thought higher taxes would berequired in order for Social Security benefits to bepaid in the next century, 73 percent respondedaffirmatively. Forty-two percent of individuals

1 See Employee Benefit Research Institute/The GallupOrganization, Inc., “Public Attitudes on Social Secu-rity, 1990,” EBRI Report G-7 (Washington, DC:Employee Benefit Research Institute, 1990); “PublicAttitudes on Social Security Benefits, 1991,” EBRIReport G-23 (Washington, DC: Employee BenefitResearch Institute, 1991); “Public Attitudes on SocialSecurity, Part I,” EBRI Report G-56 (Washington, DC:Employee Benefit Research Institute, 1993); “PublicAttitudes on Social Security, Part II,” EBRI Report G-57 (Washington, DC: Employee Benefit ResearchInstitute, 1994); and “Public Attitudes on SocialSecurity, 1995, EBRI Report G-62 (Washington, DC:Employee Benefit Research Institute, 1995).


128

0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0

40%

60%

67%

82%

Retired Americans are overly interested in expanding their Social SecurityBenefits

You expect to get less money out of Social Security than what you or your spouse put in

Most people now receiving Social Security really need the assistance provided

Working Americans are beginning to lose faith in whether Social Security benefitswill be available when they retire

Percentage

Source:Employee Benefit Research Institute/The Gallup Organization, Inc., “Public Attitudes on Social Security, 1995,” EBRI Report G-62 (Employee BenefitResearch Institute, 1995).

Appendix C: Chart 1Percentage of Respondents Who Agreed or Strongly Agreed, March 1995

surveyed in 1995 disagreed or stronglydisagreed with the statement that taxes will haveto be raised dramatically to pay for Social Securitybenefits in the future. In comparison, one-thirdagreed or strongly agreed with the precedingstatement.

Contrary to the notion that individuals do notwelcome immediate change, in the March 1995survey, Americans indicated a preference for someimmediate tax increases in order to lessen the taxburden on future workers (63 percent in favor)(appendix C, chart 2). Interviewees were informedthat, in order to maintain present levels of SocialSecurity benefits for baby boomers, the SocialSecurity payroll tax would have to increase ap-proximately 27 percent to 33 percent for bothemployers and employees by 2030. Twenty-eightpercent said they preferred to postpone taxes untilafter 2010.

■ Changes in the Level ofBenefits Received

Twenty-four percent of surveyed individuals inApril 1994 expected the level of Social Securitybenefits to increase in the future, while 40 percentexpected benefits to decrease and 31 percentbelieved they would be eliminated. Benefits wouldremain the same, according to 4 percent of respon-dents. A similar question was asked in March 1995;however, direct comparison of the responses is notpossible because the questions were phraseddifferently2 and the response options differed as

well. In March 1995, 21 percent of respondentsexpected that benefits would be reduced for allpeople, whereas 25 percent expected they would bereduced at a greater rate for higher income peoplethan for lower income people. Additionally,26 percent thought the benefits would stay thesame, and one-fifth thought they would beeliminated.

Interestingly, when individuals were askedwhat they believed should happen to the level ofbenefits (as opposed to what they expect to happen),their responses were quite different. Not surpris-ingly, a greater percentage would prefer to see onlysome people affected by reform. Five percentbelieved that Social Security benefits should bereduced for everyone, but 45 percent believed thatbenefits should be reduced more for higher incomepeople than for lower income people. Another40 percent thought benefits should stay the same,and 4 percent thought the benefits should beeliminated.

■ Expected ReturnsIn the March 1995 survey, 60 percent of respon-

2 The April 1994 question read, “Do you expect thelevel of Social Security benefits to increase, decrease, orbe eliminated in the future?” while the March 1995question read, “In the future, do you expect that SocialSecurity benefits will: (a) be reduced for all people,(b) be reduced at a greater rate for high income peoplethan for low income people, (c) stay the same, (d) beeliminated, or (e) don’t know.”

129

Appendix C: Chart 2Payroll Tax Increase Preferences,

March 1995

Source: Employee Benefit Research Institute/The Gallup Organiza-tion, Inc., “Public Attitudes on Social Security, 1995,” EBRI ReportG-62 (Employee Benefit Research Institute, 1995).

Implement some tax increases now 62% Postpone taxes

until after 2010 28%

Don’t know/refused 8%

dents supported the fact that a part of everyworking person’s income goes to support the SocialSecurity program, which is the basic premise of asocial insurance program. However, 17 percent ofindividuals were opposed to this fact. Although amajority believe that everyone should pay intoSocial Security, some believe that not everyoneshould receive benefits from the program. Thirty-two percent agreed or strongly agreed that retireeswith earnings over $100,000 should not get SocialSecurity, even if they paid into the system. How-ever, nearly one-half (47 percent) either disagreedor strongly disagreed with the previous statement.

Sixty percent of those surveyed in 1995

expected to receive less money from Social Securitythan they contributed. Interestingly, age differencesexisted for this question. Among those aged 18–34and 35–54, 72 percent and 67 percent, respectively,expected to contribute more money than they wouldreceive from Social Security. In comparison,34 percent of those aged 55 and over expected toreceive less money than they contributed.

In general, Social Security is believed to be agood program and, in 1995, 67 percent agreed orstrongly agreed that most people receiving SocialSecurity really need the assistance provided. Mostpeople are now aware of the upcoming issues facingthe program and are conscious of the need for sometype of reform.


130

131

Additional InformationAppendix

D■ Social Security Reform

BibliographyAdvisory Council on Social Security. Report of the

1994–1996 Advisory Council on Social Security.Vols. 1 and 2. Washington, DC: AdvisoryCouncil on Social Security, 1997.

Alliance Capital/Louis Harris and Associates.Survey on Financial Preparedness. New York,NY: Alliance Capital Management L.P., 1997.

Altig, David, and Gokhale, Jagadeesh. “A SimpleProposal for Privatizing Social Security.”Economic Commentary. Federal Reserve Bankof Cleveland, May 1, 1996.

American Academy of Actuaries. “Means TestingFor Social Security.” Issue Brief (Fall 1996).

___________. “Social Security Privatization: TrustFund Investment.” Issue Brief (Spring 1996).

___________. “Social Security Privatization: Indi-vidual Accounts.” Issue Brief (Spring 1996).

Atkinson, M.E., J. Creedy, and D.M. Knox. “Alter-native Retirement Income Strategies: A CohortAnalysis of Lifetime Redistribution.” EconomicRecords (June 1996): 97–106.

Baker, Dean. “The Privateers’ Free Lunch” TheAmerican Prospect (May/June 1997): 81–84.

Barreto, F.A., and O.S. Mitchell. “Privatizing LatinAmerican Retirement Systems.” BenefitsQuarterly (forthcoming).

Bartlett, Bruce, James L. Payne, and AllanCarlson. “Personal Savings Accounts WouldInvigorate the Economy and StrengthenFamilies.” The American Enterprise. Vol. 8,No. 1 (January/February 1997): 54–58.

Board of Trustees. 1997 Annual Report of the Boardof Trustees of the Federal Old-Age and Survi-vors Insurance and Disability Trust Funds.Washington, DC: U.S. Government PrintingOffice, forthcoming.

Brunner, J.K. “Transition from a Pay-As-You-Go to

a Fully Funded Pension System: The Case ofDiffering Individuals and IntragenerationalFairness.” Journal of Public Economics (April1996): 131–146.

Burtless, Gary, and Barry Bosworth. “PrivatizingSocial Security: The Troubling Trade-Offs.”Brookings Policy Brief no. 14 (March 1997).

Carter, Marshall N., and William G. Shipman.Promises to Keep: Saving Social Security’sDream. Washington, DC: Regnery Publishing,1996.

Cassidy, John. “Spooking the Boomers.” New Yorker(January 13, 1997): 30–35.

Chen, Yung-Ping. “Public and Private Approachesfor Redesigning Social Security.” Journal of theAmerican Society of CLU and ChFC (July1996): 48–52.

Committee for Economic Development. FixingSocial Security. Washington, DC: Committeefor Economic Development, 1997.

__________. Fixing Social Security: A Statement bythe Research and Policy Committee of theCommittee for Economic Development. Wash-ington, DC: Committee for Economic Develop-ment, 1997.

Contingencies. “Making Sure Its There: ActuariesDebate Social Security Reform.” Contingencies(March/April 1997): 22–32.

_________. “Privatizing Social Security.” Contingen-cies (July/August 1996): 24–30.

Corporate Benefits Institute/Corporate ResearchPanel. Social Security and Medicare in the 21st

Century: A Survey of Employee Benefits Profes-sionals from the U.S. Corporate Sector.Brookfield, WI: International Foundation ofEmployee Benefit Plans, 1996.

Davis, Trudi Jo. “The Media Monster: SwayingPublic Opinion on the State of Social Security.”Secure Retirement (January/February 1997):6–10.


132

Dentzer, Susan. “Social Security: Hot SummerSequel.” U.S. News and World Report (July 8,1996): 49.

Diamond, Peter A. “Proposals to Restructure SocialSecurity.” Journal of Economic Perspectives(Summer 1996): 67–88.

Diamond, Peter A., David C. Lindeman, andHoward Young, eds. Social Security: What Rolefor the Future? Washington, DC: NationalAcademy of Social Insurance, 1996.

Doney, Lloyd D., James P. Trebby, and George W.Kutner. “Social Security Contributions, Accu-mulations, and Payments: A Case Study.” Taxes(October 1996): 634–641.

Espina, A. “Reform of Pension Schemes in theOCED Countries.” International Labor Review.Vol. 135, No. 2 (1996): 181–206.

Feldstein, Martin. The Missing Piece in PolicyAnalysis: Social Security Reform. WorkingPaper No. 5413. Cambridge, MA: NationalBureau of Economic Research, 1996.

________. “Social Security and Saving: New TimeSeries Evidence” National Tax Journal (June1996): 151–164.

Feldstein, Martin, and Andrew Samwick. TheTransition Path in Privatizing Social Security.Working Paper No. 5761. Cambridge, MA:National Bureau of Economic Research, 1996.

Forman, Barry. “The Once and Future SocialSecurity Tax Expenditure.” Benefits Quarterly(forthcoming).

Glassman, James K., Blake Hurst, and KarlZinsmeister. “Social Security Needs a Bypass:Will Our Politicians Allow It?” The AmericanEnterprise. Vol. 8, No. 1: 59–63.

Gramlich, Edward M. “Different Approaches forDealing with Social Security.” AmericanEconomic Review (May 1996): 358–362.

Hendersen, David R. et al. “Personal SavingsAccounts Would Be Good for Everyday Work-ers.” The American Enterprise, Vol. 8, No. 1:51–53.

Hu, S.C. “Myopia and Social Security Financing.”Public Finance Quarterly (July 1996): 319–348.

International Social Security Review. “Develop-ments and Trends in Social Security,1993–1995: Report of the Secretary General,Part 1: Social Security in the 1990’s: TheImperatives of Change.” International SocialSecurity Review. Vol. 49 (February 1996): 5–16.

________. “Developments and Trends in SocialSecurity, 1993-1995: Report of the SecretaryGeneral, Part 2: Analysis of Recent Develop-ments.” International Social Security Review.Vol. 49 (February 1996): 5–16.

Johnson, P. and Stears, G. “Should the Basic StatePension Be a Contributory Benefit?” FiscalStudies (February 1996): 105–112.

Jones, Thomas W. “Strengthening the CurrentSocial Security System.” The Public Policy andAging Report. Vol. 7, No. 3 (June 1996).

________. “Maintaining Social Security.” BenefitsQuarterly (forthcoming).

Kingson, Eric R., and James H. Schulz. SocialSecurity in the 21st Century. New York, NY:Oxford University Press, 1997.

Kotlikoff, Laurence J. “Privatization of SocialSecurity: How It Works and Why It Matters.”In James Poterba, ed., Tax Policy and theEconomy 10. Cambridge, MA: National Bureauof Economic Research/MIT Press, 1996.

________. “Privatizing Social Security at Home andAbroad.” American Economic Review (May1996): 368–372.

________. Simulating the Privatization of SocialSecurity in General Equilibrium. WorkingPaper No. 5776. Cambridge, MA: NationalBureau of Economic Research, 1996.

Kritzer, Barbara E. “Privatizing Social Security:The Chilean Experience.” Social SecurityBulletin (Fall 1996): 45–55.

Langer, David “Social Fund Isn’t Really A PonziScheme of IOUs.” Pensions & Investments(June 24, 1996): 4–5.

Luciano, Lani. “Support Builds for Saving theSocial Security System by Investing in Stocks.”Money (January 1997): 20–22.

Mercer. “The Future of Social Security.” The MercerReport (January 24, 1997): 1–4.

Mitchell, Olivia S., and Stephen P. Zeldes. SocialSecurity Privatization: A Structure for Analysis.Working Paper No. 5512. Cambridge, MA:National Bureau of Economic Research, 1996.

Myers, E.J. Let’s Get Rid of Social Security: HowAmericans Can Take Charge of Their OwnFuture. Amherst, NY: Prometheus Books, 1996.

Myers, Robert J. “Privatization of Social Security: AGood Idea?” Journal of the American Society ofCLU and ChFC (July 1996): 42–46.

________. “The Social Security Sky Is Not Falling.”

133

Challenge (November-December 1996): 23–24.National Association for the Self-Employed/Small

Business Research Institute. The NASE Post-Election National Survey, January 7–9, 1997.Hurst, TX: NASE, January 18, 1997.

Ostuw, Pamela “Attitudes on Social SecurityReform,” EBRI Notes, No. 3 (Employee BenefitResearch Institute, March 1977): 3–5.

Paine, T.H. “Appraisal of Alternative Ways to FixSocial Security.” Benefits Quarterly,(forthcoming).

Panis, Constantijn W.A., and Lee A. Lillard.Socioeconomic Differentials in the Returns toSocial Security. Santa Monica, CA: RANDCorporation, 1996.

Payne, James L., et al. “Social Security Is Broken.”The American Enterprise. Vol. 8, No. 1: 41–44.

Petersen, Peter G. Will America Grow Up Before ItGrows Old? New York: Random House, Inc.,1977.

Pianin, Eric, and Brossard, Mario. “AmericansOppose Cutting Entitlements to Fix Budget.”Washington Post, 29 March 1997: A4.

Pinera, Jose. “Empowering Workers: ThePrivatization of Social Security in Chile.” CatoJournal (Fall–Winter 1995-1996): 155–166.

Quadagno, Jill. “Social Security and the Myth ofthe Entitlement ‘Crisis’.” The Gerontologist.Vol. 36, No. 3 (1996): 391–399.

Quinn, Joseph F. Entitlements and the FederalBudget: Securing Our Future. Washington, DC:National Academy on Aging, 1996.

Rehfeld, Barry. “Fixing Social Security.” Institu-tional Investor (December 1996): 82–91.

Reichlin, P., and P. Siconolf. “The Role of SocialSecurity in An Economic with AsymmetricInformation and Financial Intermediaries.”Journal of Public Economics (May 1996):153–175.

Reimers, C., and M. Honig. “Responses to SocialSecurity by Men and Women: Myopic and Far-Sighted Behavior.” Journal of Human Re-sources (Spring 1996): 359–382.

Rivers, W. A. “Will the Social Security SystemContinue to Offer Social Security?” BenefitsQuarterly (forthcomng).

Robertson, A. Haeworth. The Big Lie: What EveryBaby Boomer Should Know About SocialSecurity and Medicare. Washington DC:Retirement Policy Institute, 1977.

Rosati, F.C. “Social Security in a Non-altruisticModel with Uncertainty and EndogenousFertility.” Journal of Public Economics (May1996): 283–294.

Ross, Stanford G. “Public Versus Private Pensions:Dimensions of a World-Wide Debate.” InProtecting Retirement Incomes: Options forReform. Geneva: International Social SecurityAssociation, 1996.

Salisbury, Dallas L. “Reforming Social Security andOther Public Entitlements: Should EmployersCare?” ACA News (May 1996): 6–9.

Salisbury, Dallas L., and Olsen, Kelly A. “SocialSecurity: Will the New Advisory Council ReportLead to Action?” ACA News (February 1997):10–14.

Schieber, Sylvester, J. “A New Vision for SocialSecurity.” The Public Policy and Aging Report.Vol. 7, No. 3 (June 1996).

Schieber, S.J., and J. B. Shoven. “Social SecurityReform: Around the World in 80 Ways.” Ameri-can Economic Review (May 1996): 373–377.

Social Security Administration. “Actuarial Status ofthe Social Security and Medicare Trust Funds.”Social Security Bulletin (Summer 1996): 57–63.

Stern, Linda. “Can We Save Social Security?”Modern Maturity (January/February 1997):28–36.

Steuerle, Gene. “Privatizing Social Security: AThird Option.” Tax Notes (Dec. 16, 1996):1357–1358.

Sullivan, Carl. “Social Security Reform GainsMomentum.” Financial Planning (April 1996):43–48.

Sweeney, J. America Needs a Raise: Fighting forEconomic Security and Social Justice. Boston,MA: Houghton Mifflin, 1996.

Tanner, Michael. “Privatizing Social Security: A BigBoost for the Poor.” The Cato Project on SocialSecurity Privatization. SSP no. 4. Washington,DC: Cato Institute, July 26, 1996.

________. “Public Opinion and Social SecurityPrivatization.” The Cato Project on SocialSecurity Privatization. SSP no. 5. Washington,DC: Cato Institute, August 6, 1996.

Taylor, Jeff A. “Mixed Signals on Social Security:Public Sees Problem But Resists Changes, forNow.” Investors Business. 4 March 1997, A1.

Third Millennium. “Americans Ready for MajorChanges in Social Security, Third Millennium


134

Poll Finds.” News Release, December 20, 1996.________. Freed From FICA. New York, NY: Third

Millennium, 1997.Twentieth Century Fund. Social Security Reform:

The Basics. New York, NY: The TwentiethCentury Fund, 1996.

U.S. Congress. Congressional Budget Office. Long-Term Budgetary Pressures and Policy Options.Washington, DC: Congressional Budget Office,1997.

U.S. Congress. Senate. Special Committee onAging. Social Security Reform Options: Prepar-ing for the 21st Century. S. Hrg. 104–705.Washington, DC: U.S. Government PrintingOffice, 1996.

U.S. Congress. Senate. Subcommittee on Aging ofthe Committee on Labor and Human Re-sources. Ensuring Our Workers’ RetirementSecurity: Proposals for Reform. S. Hrg.104–523. Washington, DC: U.S. GovernmentPrinting Office, 1996.

U.S. Congress. Staff of the House Committee onWays and Means. Overview of EntitlementPrograms: 1996 Green Book: BackgroundMaterial and Data on Programs Within theJurisdiction of the Committee on Ways andMeans. 104th Congress, Second Session.Washington, DC: U.S. Government PrintingOffice, 1996.

U.S. Social Security Administration. Compilation ofthe Social Security Laws: Including the SocialSecurity Act as Amended, and Related Enact-ments through January 1, 1995. 2 Vols.,Biennial. Washington, DC: U.S. GovernmentPrinting Office, 1995.

U.S. Social Security Advisory Board. DevelopingSocial Security Policy: How the Social SecurityAdministration Can Provide Greater PolicyLeadership. Washington, DC: Social SecurityAdvisory Board, March 1997.

Walker, David M. Retirement Security: Understand-

ing and Planning Your Financial Future. NewYork, NY: John Wiley & Sons, 1997.

Warshawsky, Mark J. “The Market for IndividualAnnuities and the Reform of Social Security.”Benefits Quarterly (forthcoming).

Weaver, Carolyn, Joel Kotkin, and Karlyn Bowman.“There Is an Alternative.” The AmericanEnterprise. Vol. 8, No. 1: 44–50.

Weinberger, Mark. “Social Security: Facing theFacts.” The Cato Project on Social SecurityPrivatization. SSP no. 3. Washington, DC: CatoInstitute, 1996.

Wheeler, Peter M., and John R. Kearney. “IncomeProtection for the Aged in the 21st Century: AFramework to Help Inform the Debate.” SocialSecurity Bulletin (Summer 1996): 3–199.

White, Lawrence J. “Investing the Assets of theSocial Security Trust Funds in Equity Securi-ties: An Analysis.” Investment CompanyInstitute Perspective (May 1996).

White, Stephen A. “Are the Forecasts Too Good toBe True? Social Security Privatization.” Per-spectives. Milliman & Robertson (November1996): 1–3.

Wiedmer, T. “Growth and Social Security.” Journalof the Institute of Theoretical Economics(September 1996: 531–539.

Wilcox, Melinda Dovel. “Saving Social Security.”Kiplinger’s Personal Finance Magazine (April1997): 61–68.

Yakoboski, Paul. “The Reality of Retirement Today:Lessons in Planning for Tomorrow.” EBRI IssueBrief no. 181 (Employee Benefit ResearchInstitute, January 1997).

Young America’s Foundation. “Privatizing SocialSecurity.” Issue Briefs. Young America’s Foun-dation, 1996.

135

■ Internet Sites with Informationon Social Security Reform

1994-1996 Advisory Council on Social Securityhttp://www.ssa.gov/policy/adcouncil/toc.htm

American Council for Capital Formation, ACCFCenter for Policy Researchhttp://www.accf.org

The Brookings Institutionhttp://www.brook.edu

The Cato Institutehttp://www.cato.org

Citizens for a Sound Economyhttp://www.cse.org/cse/welcome.html

Committee for Economic Development (CED)http://ced.sohonet.com

Common Sense for America (CSFA)http://www.csfa.org

The Democratic Leadership Council & ProgressivePolicy Institutehttp://www.dlcppi.org

The Electronic Policy Networkhttp://www.epn.org

Employee Benefit Research Institutehttp://www.ebri.org

Institute for Economic Analysishttp://cpcug.org/user/jatlee

IntellectualCapital.Comhttp://www.intellectualcapital.com

The Maxwell School Center for Policy Research(CPR)http://www-cpr.maxwell.syr.edu

National Academy of Social Insurancehttp://www.idsonline.com/nasi

National Association of Manufacturershttp://www.nam.org

National Bureau of Economic Research, Inc.http://www.nber.org

National Commission on Retirement Policyhttp://www.csis.org/retire/

Pension Research Councilhttp://rider.wharton.upenn.edu/~mitchelo/prc.html

Policy.comhttp://www.policy.com

The Progress & Freedom Foundationhttp://www.pff.org/pff/

RANDhttp://www.rand.org

Social Security Administrationhttp://www.ssa.gov

The Social Security Crisishttp://www.panix.com/~stern/ss/socindex.html

The Social Security Reform Research Pagehttp://www.execpc.com/~freitag/reform.htm

The Tax Foundationhttp://www.taxfoundation.org/index.html

The Urban Institutehttp://www.urban.org


136

137

Appendix

Policy Forum Attendees List

Lawrence AtkinsThe Jefferson Group

David BaerAARP

Robert BallFormer Commissioner, Social SecurityAdministration

Laurel BeedonAARP

William BeemanCommittee for Economic Development

James BellMead Corporation

Ruth BlackerAARP

David BlitzsteinUnited Food & Commercial Workers Int’l Union

Chris BoneActuarial Sciences Associates

Francis BonsignoreMarsh & McLennan

Hugh BradyBellSouth Corporation

Judith BurnsDow Jones News Service

Gary BurtlessThe Brookings Institution

David CertnerAARP

David ChandlerAtlantic Richfield Company

William Chapman, IIKemper Retirement Plan Group

William CheneyJohn Hancock

Judy ChesserSocial Security Administration

Elaine ChurchPrice Waterhouse LLP

Lee CohenAARP

Susan ColburnSBC Communications, Inc.

Geri ColombaroIBM Corporation

Ann CombsWilliam M. Mercer Co.

Christopher ConteEBRI Fellow

Charles CookCook Political Report

George CowlesBankers Trust Company

Sandy CrankSocial Security Administration

Marcy CrequeAARP

Susan CrownCitibank, N.A.

Paul CullinanCongressional Budget Office

Patsy D’AmelioEBRI

John DayCIGNA Corporation

E


138

James DelaplaneOffice of U.S. Representative Pomeroy

Susan DentzerUS News & World Report

Terri DevineCongressional Budget Office

Richard DunnGeneral Electric Company

Michelle EhmBarclays Global Investors

Bill EnglishCouncil of Economic Advisors

Lou EnoffEnoff Associates

Stephen EntinIRET

Kimberly FamigliettiZurich Kemper Investments, Inc.

Gus FaucherU.S. Department of Treasury

John FeldtmoseFoster Higgins

Mark FettingPrudential Securities

Edith FierstFierst & Moss, P.C.

Jill FinsenAARP

Michael FlahermanCalPERS

Howard FluhrThe Segal Company

Robert FriedlandNational Academy on Aging

Ed FriendEFI Actuaries

Paul FronstinEBRI

Russ GalerInvestment Company Institute

Bob GarretsonGeorge Washington University

Ron GebhardtsbauerAmerican Academy of Actuaries

John GistAARP

Robert GlennNational Education Association

Leonard GlynnState Street Bank & Trust

Steve GossSocial Security Administration

Dot GramesAARP

Peter GrayCiticorp

Janice GregoryThe ERISA Industry Committee

Dave GustafsonPBGC

Nick GwinOffice of U.S. Representative Kennelly

Eiji HabuEmbassy of Japan

Randy HardockDavis & Harman

Mark HarfIBM Corporation

Terri HarrisonState Teachers Retirement System of Ohio

Paul HewittNational Taxpayers Union Foundation

Richard HinzU.S. Department of Labor

Donald HoffmeyerSocial Security Administration

Sarah HoldenFederal Reserve Board

Martin HolmerPolicy Simulation Group

139

Neil HoweNational Taxpayers Union Foundation

Richard HubbardArnold & Porter

Thomas HungerfordU.S. General Accounting Office

Kenneth HutchinsonHigher Education Benefit Research Group

Bill IuloAARP

Richard JacksonNational Taxpayers Union Foundation

Lynn JacobsCIGNA Corporation

Melissa KahnAmerican Council of Life Insurance

Michael KahnNational Education Association

Connie KainCitibank, N.A.

David KempsU.S. Chamber of Commerce

Daniel KohlerICMA Retirement Corporation

Geoff KollmannCongressional Research Service

David LangerDavid Langer Co., Inc.

Jane LassnerJP Morgan

Gene LehrmannAARP

William Lessard, Jr.National Health Policy Forum

Dudley LesserAARP

Jules LichtensteinAARP

David LindemanWorld Bank

Brendan LynchThe Travelers Group

Margaret MaloneSocial Security Advisory Board

Joyce ManchesterCongressional Budget Office

Kenneth MannellaSocial Security Administration

Merrill MatthewsNational Center for Policy Analysis

Judy MazoThe Segal Company

Peter McCauleyPharmacia & Upjohn

John McCormackTIAA-CREF

Ken McDonnellEBRI

Lauch McLeanAARP

Peter MerrillPrice Waterhouse LLP

Curt MikkelsenEBRI Fellow

Girard MillerICMA Retirement Corporation

Tom MillerCompetitive Enterprise Institute

Deborah MilneEBRI

Olivia MitchellThe Wharton School

Jim MobergPacific Telesis Group

Marilyn MoonThe Urban Institute

Evelyn MortonAARP

Frank MulveyU.S. General Accounting Office


140

Robert MyersActuarial Consultant

Dave NadigBarclays Global Investors

Mary Jane O’GaraAARP

Kelly OlsenEBRI

Pamela OstuwEBRI

Thomas PaineEBRI Fellow

Carolyn PembertonEBRI

Joe PerkinsAARP

Martha PhillipsConcord Coalition

Bill PierronEBRI

Richard PreyThe Principal Financial Group

Helene RayderMutual of Omaha

Kenneth ReifertMerrill Lynch & Co., Inc.

Virginia RenoNational Academy of Social Insurance

Robert ReynoldsFidelity Investments-FIRSCO

Jean Marie RickettsSocial Security Administration

Sara RixAARP

Carole RobertsTravelers Group

Steve RobinsonHouse Majority Whip

Melvyn RodriguesEBRI Fellow

Mary Ross(retired) Social Security Administration

Stanford RossArnold & Porter

Joseph RusbarskyOppenheimer Capital

Norihiko SakamotoThe Brookings Institution

Dallas SalisburyEBRI

Don SauvigneIBM Corporation

Richard SawayaAtlantic Richfield Company

Tom SchlossbergDiversified Investment Advisors

Brigitte SchmidtOffice of U.S. Representative Colby

Dan SchulderNational Council of Senior Citizens

Mary ScottAARP

Bert SeidmanNational Council of Senior Citizens

John SeiterCapital Guardian Trust Company

Robert ShapiroProgressive Policy Institute

Jennifer SharpFederal Reserve Board

David SkovronKwasha Lipton

James SmithThe RAND Corporation

Marie SmithAARP

Chris SoaresProgressive Policy Institute

Robert Sollman, Jr.MetLife Insurance Company

141

Judy SoltzCIGNA Corporation

Michael SternInvestment Company Institute

Eugene SteuerleThe Urban Institute

Carolyn StewartEBRI

Kenichi TanakaEBRI Fellow

Richard ThauThird Millenium

Lawrence ThompsonThe Urban Institute

Mary TuckerAARP

Marc TwinneySocial Security Advisory Council

Jack VanDerheiTemple University and EBRI Fellow

Dan VinodAT&T

David WalkerArthur Andersen LLP

Robert WardenOPPOSE Group

Mark WarshawskyTIAA-CREF

Ross WebbAARP

Ke Bin WeiAARP

Anne WilletteUSA Today

Rudolph WilsonAARP

Stan WisniewskiNational Education Association

Paul YakoboskiEBRI

ASSESSING SOCIAL SECURITY REFORM ALTERNATIVES · i EBRI EMPLOYEE BENEFIT RESEARCH INSTITUTE®...

Documents

Transcript of ASSESSING SOCIAL SECURITY REFORM ALTERNATIVES · i EBRI EMPLOYEE BENEFIT RESEARCH INSTITUTE®...