Pension Task Force report on public employee retirement systems

MISSING IMAGE

Material Information

Title:
Pension Task Force report on public employee retirement systems
Uncontrolled:
Public employee retirement systems
Physical Description:
v, 858 p. : ; 24 cm.
Language:
English
Creator:
United States -- Congress. -- House. -- Committee on Education and Labor. -- Subcommittee on Labor Standards. -- Pension Task Force
United States -- Congress. -- House. -- Committee on Education and Labor
Publisher:
U.S. Govt. Print. Off.
Place of Publication:
Washington
Publication Date:

Subjects

Subjects / Keywords:
Local officials and employees -- Pensions -- United States   ( lcsh )
State governments -- Officials and employees -- Pensions   ( lcsh )
Civil service -- Pensions -- United States   ( lcsh )
Officials and employees -- Pensions -- United States   ( lcsh )
Genre:
federal government publication   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Bibliography:
"Public pension plans, a bibliography" p. 769-858.
Additional Physical Form:
Also available in electronic format.
Statement of Responsibility:
Committee on Education and Labor, House of Representatives, Ninety-fifth Congress, second session.
General Note:
At head of title: 95th Congress, 2d session. Committee print.
General Note:
Issued Mar. 15, 1978.
General Note:
Reuse of record except for individual research requires license from Congressional Information Service, Inc.
General Note:
CIS Microfiche Accession Numbers: CIS 78 H342-12
General Note:
Reuse of record except for individual research requires license from LexisNexis Academic & Library Solutions.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 022472141
oclc - 04157542X
Classification:
lcc - KF49
System ID:
AA00025895:00001

Full Text








J
2.4












4l4
NSIO TAS FOCE RPOR ON PUBLIC MPLOEE RTIREENT SYSTEMS




COM3TTEON EUCAIONAND LABOR ..NI FITH CNGRSS









404





V4,;









95th Congress Co TEE PRINT
2d Session









PENSION TASK FORCE REPORT ON PUBLIC

EMPLOYEE RETIREMENT SYSTEMS






COMMITTEE ON EDUCATION AND LABOR HOUSE OF REPRESENTATIVES

NINETY-FIFTH CONGRESS
SECOND SESSION














MARCH 15t 1978









Printed for the use of the Committee on Education and Labor


U.S. GOVERNMENT PRINTING OFFICE 74-us WASHINGTON : 1978


For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402














COMMITTEE ON EDUCATION AND LABOR

CARL D. PERKINS, Kentucky, Chairman FRANK THOMPSON, JR., New Jersey ALBERT H. QUIE, Minnesota
JOHN H. DENT, Pennsylvania JOHN M. ASHBROOK, Ohio
JOHN BRADEMAS, Indiana JOHN N. ERLENBORN, Illinois
AUGUSTUS F. HAWKINS, California RONALD A. SARASIN, Connecticut
WILLIAM D. FORD, Michigan JOHN BUCHANAN, Alabama
PHILLIP BURTON, California JAMES M. JEFFORDS, Vermont
JOSEPH M. GAYDOS, Pennsylvania LARRY PRESSLER, South Dakota
WILLIAM "BILL" CLAY, Missouri WILLIAM F. GOODLING, Pennsylvania
MARIO BIAGGI, New York BUD SHUSTER, Pennsylvania
IKE ANDREWS, North Carolina SHIRLEY N. PETTIS, California
MICHAEL T. BLOUIN, Iowa CARL D. PURSELL, Michigan
ROBERT J. CORNELL, Wisconsin MICKEY EDWARDS, Oklahoma
PAUL SIMON, Illinois EDWARD P. BEARD, Rhode Island LEO C. ZEFERETTI, New York GEORGE MILLER, California RONALD M. MOTTL, Ohio MICHAEL 0. MYERS, Pennsylvania AUSTIN J. MURPHY, Pennsylvania JOSEPH A. LE FANTE, New Jersey TED WEISS, New York CEC HEFTEL, Hawaii BALTASAR CORRADA, Puerto Rico DALE E. KILDEE, Michigan

SUBCOMMITTEE ON LABOR STANDARDS

JOHN H. DENT, Pennsylvania, Chairman PHILLIP BURTON, California JOHN N. ERLENBORN, Illinois
JOSEPH M. GAYDOS, Pennsylvania JOHN M. ASHBROOK, Ohio
WILLIAM "BILL" CLAY, Missouri RONALD A. SARASIN, Connecticut
MARIO BIAGGI, New York MICKEY EDWARDS, Oklahoma
LEO C. ZEFERETTI, New York ALBERT H. QUIE, Minnesota
MICHAEL 0. MYERS, Pennsylvania Ex Officio
AUSTIN J. MURPHY, Pennsylvania BALTASAR CORRADA, Puerto Rico PAUL SIMON, Illinois GEORGE MILLER, California CARL I). PERKINS, Kentucky, Jx Officio
(II)














,TABLE- OF CONTENTS

Page
Part I-Committee findings and 1
Part II-Fe'deral law presently affecting public employee retirement
systems ---------------------------------------------------------- 7
14th Amendment-Due process --- 7
-----------------------------14th Amendment-Equal protection ----------------------------- 13
11th 16
Bases of Federal ------------ 17
Civil Rights Act of 1964-Title 22
Age Discrimination in Employment Act- ---------------------- 25
Securities acts ------------------------------------------------- 28
Internal Revenue 30
Social Security -----------------------------------------------Revenue 36
Fair Labor Standards 37
Labor Management Relations Act -------------- : ----------------- 38
Comprehensive Employment and Training 39
Bankruptcy Act ----------------------------------------------- 39
Military Selective Service 40
Cost Accounting Standards Board ------------------------------- 42
Part III-State laws applicable to public employee retirement systems- - 43 Part IV-Characteristics and operations of public employee retirement 47
A. Introduction
Extensive studies 47
PERS influence vast and diversity great 48
Technical note -------------------------------------------- 49
Acknowledgment ------------------------------------------ 49
B. PERS Universe- General characteristics
Number and 51
Defined benefit and defined contribution---------------------- 53
Insured 54
Contributory and noncontributory--------------------------- 54
Plans by jurisdiction --------------------------------------- 55
Administration and coverage categories----------------------- 56
Single and multiple employer plans -------------------------- 58
Social Security coverage ------------------------------------ 58
Pension coverage not universal ------------------------------ 59
C. PERS Administration, reporting, and disclosure
PERS development and consolidation ------------------------ 61
Public pension 63
Retirement system administration --------------------------- 65
Retirement system 69
Plan disclosure to 71
Internal Revenue Code 77
Recordkeeping and other 79
Summary and 79
D. PERS participation, vesting, portability, and plan termination
provisions
Participation (membership) requirements --------------------- 83
Vesting and related requirements ---------------------------- 87
Portability 92
Risk of public pension benefit loss or reduction---------------- 95
Summary and 99
(M)






TV

Part TV-Continued
E. PERS benefit structure and income replacement levels Page
Retirement ag----------------------103
Disability benefits ----------------------------------------- 107
Pre-retirement death bnft----------------107
Post-retirement death bnft----------------108
Post-retirement cost-of-living adjustments ----------------------108
Retirement benefit fomls----------------109
Gross income replacement rae---------------115
Net income replacement rates --------------------------------119
Summary and conclusions -----------------------------------126
F. PERS finances
Assets and icm---------------------129
Employee and employer contributions and benefit payments---.- 135 Employee contribution rts----------------138
Source of employer contributions-----------------------------139
G. PERS funding
Need for reserve fudn------------------144
Present funding practices----------------------------------- 150
Actuaries, actuarial valuations and assumptions -----------------158
Current funding status of defined benefit pln---------163 PERS unfunded accrued liability -----------------------------163
"Quick Liability" as a measure of PERS funding ----------------166
Reserve Ratio as a measure of PERS funding------------------168
Assets to benefit payments ratio as a measure of PERS funding --. 170 Comparison of public and private pension systems ---------------172
Summary and cnlios-----------------179
11 PERS, fiduciary and investment practices
Discosue------------------------------------183
Awareness of legal poiin----------------184
Enforcement of plan prvsos---------------185
Accounting, auditing, and actuarial standards -------------------185
Corruption and dsoet-----------------186
Recordkeeping and claims procedures -------------------------187
Dilienc-------------------------------------187
Plan assets and fiduciary responsibility------------------------188
Local inetet---------------------190
Substantive standards --------------------------192
Management and investment lmttos-----------194
Summary and cnlsos-----------------195











APPENDICES
I. Tabulax results of the Pension Task Force survey of public employee Page
retirement systems ----------------------------------------- 199
II. Pension Task Force survey questionnaire ----------------------- 301
III. Technical note on the Pension Task Force survey methodology---- 323 IV. Number and membership of public employee retirement systems --- 325
V. Summary and analysis of the pension and retirement systems of
the 50 States and the District of 397
VI. The governmental pension and retirement system of the State of
Hawaii: Coverage, funding, financing and fiduciary standards- 473 VII. State and local pension plans for the State of Illinois: Coverage, funding, financing, and fiduciary 481
VIII. Governmental pension plans of the State of New York and the
city of New York: Coverage, funding, financing, and fiduciary
standards ------------------------------------------------- 585
IX. Interrelationship of retirement programs covering employees of
Federal agencies and instrumentalities------------------------ 663
X. Letter (August 15, 1977) from Internal Revenue Service responding
to questions concerning the tax qualification of public employee
retirement 673
XI. Letter (October 7, 1977) from Internal Revenue Service responding
to questions concerning the tax status of Federal retirement
systems -------------------------------------------------- 677
XII. GAO Report of August 3, 1977: Federal retirement systems: Unrecognized costs, inadequate funding, inconsistent benefits------ 689 XIII. Department of Labor bibliography (May 1977): Public pension 769
XIV. Twentieth Century Fund Report: Conflicts of Interest: State and
Local Pension Fund Asset Management, by Louis M. Kohlmeier- 859



















Digitized by the Internet Archive
in 2013'













http://archive.org/detaiIs/pensioceOOunit











PART I-COMMITTEE FINDINGS AND CONCLUSIONS
Pursuant to section 3031 of Public Law 93-406, The Employee Retirement Income Security Act of 1974 (ERISA), the Committee on Education and Labor and the Subcommittee on Labor Standards of the House of Representatives have undertaken a comprehensive study of federal, state, and local public employee retirement systems.
Section 3031 of ERISA requires that:
The Committee on Education and Labor and the Committee
on Ways and Means of the House of Representatives and the Committee on Finance and the Committee on Labor and Public Welfare of the Senate shall study retirement plans established and maintained or financed directly or indirectly) by the Government of the United States, by any State (including the District of Columbia) or political subdivision thereof, or by any agency or instrumentality of any of the foregoing. Such study shall include
an analysis of(1) the adequacy of existing levels of participation, vesting
and financing arrangements,
(2) existing fiduciary standards, and
(3) the necessity for Federal legislation and standards with
respect to such plans.
In determinm*gr whether any such plan is adequately financed, each committee shall consider the necessity for minimum funding standards, as well as the taxing power of the government maintaining the pian.
The CongFess,--in 1974, excluded governmental retirement systems from the major provisions of ERISA in order that additional information might be obtained regarding whether a need exists for further regulation of govern ntal plans. This report is intended to provide information to the Congress in accordance with the ERISA mandate.
The information and conclusions in this report are based on extensive research and investigations conducted by the Committee. Public hearings relating to the retirement systems covering state and local government employees and the need for federal standards regulating such systems were conducted during 1975 in Washington, D.C., California, Connecticut, and Illinois. In addition, numerous meetings were held with representatives of every element of the governmental pension plan community. A comprehensive survey f governmental plans was conducted, covering 96% of all public pension plan participants. Virtually every element of the design and operation of governmental pension plans was explored in order -to fully document the universe of public plans and their characteristics. In addition, extensive research was undertaken to identify and record the various federal, state, and local laws affecting the design and maintenance of public employee retirement systems.
On the basis of the facts and information which have been received and studied, your Committee makes the following conclusions:
(1)




2

IN GENERAL
The universe of public employee retirement systems (PERS) exerts a substantial influence on the economic, social, and political fabric of the United States. The far reaching influence of the PERS involves a fundamental national interest affecting the well-being and security of millions of workers and their families, the operation of the national economy, the revenues of the United States, and the relationships between the federal government and state and local governments. Recent experience shows the growth in the size and scope of the PERS to be rapid, substantial, and increasingly interstate. The national securities and other capital formation markets are heavily influenced by the investment of over $115 billion in assets that are currently held by the public employee retirement systems of state and local governments. The manner in which the assets of state and local government retirement systems are invested in the future will have a direct effect on the well-being and economic security of the 13 million current Participants as well as the future participants in such plans. The provisions and significance of the PERS have not been fully comprehended by plan participants, plan officials, other government officials, and taxpayers generally. As a result, the current regulatory framework applicable to the PERS does not adequately protect the vital national interests which are involved.

EXISTING FEDERAL REGULATION
A number of federal constitutional and statutory provisions affect the PERS in a significant manner. In many instances the precise impact that these laws have on governmental plans is not yet clear. The most significant body of federal law presently affecting the PERS is found in sections 401 and 503 of the Internal Revenue Code (I RC). However, the ambiguity of the I RC provisions as they relate to the PERS and the inconsistent interpretation and enforcement by the Internal Revenue Service of the non-discrimination and other plan qualification requirements have created confusion and sharply limited the protections such provisions offer to plan participants. The absence of any single federal agency to coordinate the administration and enforcement of the various federal laws relating to retirement income has p~recludedl the development of a unified national policy with regard to either public employees retirement systems or private pension plans.
EXISTING STATE REGULATION
The various state constitutional and statutory provisions relating to state and local government retirement systems do not as a whole provide a uniform and adIequate means of protecting the interests of plan participants in their retirement systems. Statutory provisions diirectly relating to the establishment and operation of public employee retirement systems are often non-existent. In cases where relevant state laws (10 exist, thiey are often found to be ambiguous, thus creating legal uinc.ertainties which are seldom clarified by the courts. In those instances in wich state l.aws relating to governmental pension Systems hiave been judicially clarified1, the courts have frequently interpreted the statutory or constitutional provisions in a manner that renders






3

them far less -protective of participant interests than might be anticipated from a plain reading of such laws. Generally, the states have failed to establish (1) clear standards regulating the activities of plan fiduciaries, and (2) effective remedies for plans and plan participants when injury occurs as a result of abuse. The absence of a coherent and uniform regulatory framework has resulted in generally ineffective communication of basic plan provisions, inadequate safeguarding of plan assets and insufficient protection of participants' interests.

REPORTING AND DISCLOSURE OF PLAN PROVISIONS AND
FINANCIAL AND ACTUARIAL INFORMATION
Serious deficiencies exist among public employee retirement systems at all levels of government regarding the extent to which important information is reported and disclosed to plan participants, public officials, and taxpayers. In many cases plan participants are not informed of basic plan provisions or the amount of their vested benefits, thus leading to unwarranted forfeitures of earned benefits. Public employee retirement systems at all levels of government are not operated in accordance with the generally accepted financial and accounting procedures applicable to private pension plains and other important financial enterprises. The potential for abuse is great due to the lack of independent an d external reviews of the operations of many plans. There is an incomplete assessment of true pension costs at all levels of government due to the lack of adequate actuarial valuations and standards. In general, the absence of uniform accounting, auditing, and actuarial standards impeaches the credibility of the current practices in these areas.

PARTICIPATION, VESTING AND OTHER BENEFIT
P ROVISIONS, AND PLAN TERMINATIONS
Generally, the benefit levels and benefit provisions of public employee retirement systems compare favorably with those found under private sector pension systems. For example, the plans covering only a very small minority of all public employees fail to meet ERISA's minimum participation and benefit accrual requirements. However, 70 percent of the plans, covering about one-fifth of all state and local government employees, fail to mee t ERISA's minimum vesting requirements (as do the federal plans covering the uniformed services). A majority of all public plans (including the Federal Civil Service Retirement System) also fail to meet an ERISA-like minimum requirement that 5 percent interest be paid upon the return to a terminated participant of all mandatory employee contributions. In like manner, nearly all public pension plan participants lack ERISA's protection against the forfeiture of the employer financed portion of vested benefits upon the withdrawal of employee mandatory contributions. A substantial number of public employees at the state and local level could benefit from ERISA's provisions (1) requiring joint and survivor annuities, and (2) prohibiting restrictive "break-in-service" rules. With regard to benefit levels,, a small number of public pension plans were found to violate the maximum benefit limitations of section






4

415 of the Internal Revenue Code. Plan terminations and insolvencies are rarely found in the PERS. In the few instances in which plan terminations and insolvencies have occurred, participants suffered temporary, but no permanent, benefit losses.

FUNDING
In the vast majority of public employee retirement systems, plan participants, plan sponsors, and the general public are kept in the dark with regard to a realistic assessment of true pension costs. The high degree of pension cost blindness which pervades the PERS is due to the lack of actuarial valuations, the use of unrealistic actuarial assumptions, and the general absence of actuarial standards. Nonetheless, a significant minority of public plans appear to have accumulated substantial pension reserves through the use of ERISA-like actuarial funding methods. However, the majority of public plans are experiencing rising pension costs as a percentage of payroll due to the lack of actuarial funding practices. Approximately 17 percent of governmental plans continue to use the discredited pay-as-you-go financing approach to satisfy benefit obligations. Efforts should be made to eliminate any incentives which mav exist for public plans to continue using such inappropriate financing methods. Efforts should also be made to encourage the accumulation of pension reserves through the use of actuarial funding methods.
Furthermore, adherence by governmental retirement systems to a proper funding policy, in those plans in which such a policy exists, is often hindered by the absence of a rational relationship between the source and level of plan revenues and the funding needs of the plan. Pension plan revenues may lack stability and predictability due to the indeterminate amount of funds available from stipulated allocations of state insurance premium taxes, federal revenue sharing funds, or limited special tax levies. Given the increasingly restricted revenues for many public pension plans, there is a compelling need for uniform actuarial measures, terminology, and standards to enable plan participants, plan sponsors, and taxpayers to assess the present funding status and future funding needs of their systems.
PLAN CONTROL AND FIDUCIARY RESPONSIBILITY
Control over the administration of state and local public employee retirement systems, and the management and investment of public plan assets, is frequently inadequate. Often, public plans are administered without the benefit of either statutory guidance or even written plan documents, creating open opportunities for abuse. At the other extreme, public pension plans may be enveloped in a maze of laws leading to conflicts, confusion, and the inadequate allocation of fiduciary responsibilities. The absence of a uniform standard of conduct applicable to plan fiduciaries inhibits the proper discharge by fi]uciaries of their responsibilities, thereby hindering the proper operation of public employee retirement systems. Past breakdowns in the accountability of plan fiduciaries have led to favoritism and abuse in benefit determinations, failure to disclose information vital to the interests of plan participants. violations of the Internal Revenue





5

Code, and even pension plan insolvencies. There is presently no uniform requirement that public pension plan fiduciaries manage and invest pension plan assets prudently and for the exclusive benefit of plan participants and beneficiaries or that such fiduciaries be held responsible for any loss resulting from a breach of their fiduciary duties. Because adequate safeguards have not been erected, conflicts of interests in many instances have been permitted to ripen into clear examples. of fiduciary abuse, with resultant losses of I pension plan assets and income. Additionally, the investment of public pension. funds in securities issued by state and local governments is, in most cases, inappropriate in light of the low-yielding, non-taxable nature of such securities and the tax-exempt nature of governmental plans. Because the investment of public pension funds in state and local government securities is fraught with conflict and great potential for the impairment of pension plan stability, it may be appropriate in such circumstances to limit such investments. Numerous restrictive investment practices imposed by statute or policy have also served to impair public pension plan investment returns. Clearly, a uniform standard of fiduciary conduct is neede& to conform public employee retirement system administrative and investment practices with the practices expected of other important financial enterprises.















-PART II-FEDERAL LAW PRESENTLY AFFECTING PERS,
In consideration of the need for federal legislation directly regulating state and local governmental retirement systems, it is frequently overlooked that a large number of federal laws presently regulate various aspects of public employee retirement systems. An understanding of this existing federal regulation is vital to an appreciation of whether additional federal legislation is needed in this area, and if so, what course such legislation should take in relation to the existing law. This Part of the Report is intended to provide an overview of the present regulatory environment. As discussed below, the present impact of various federal laws on governmental plans is in many instances still in its formative stages. As awareness of the existing federal regulation grows, and the enforcement of these various federal laws is improved, it is clear that the federal laws already in existence will have an increasingly greater influence on the design and operation of state and local government pension plans.

FOURTEENTH AMENDMENT-DUE PROCESS
The Due Process Clause of the 14th Amendment1I is highly relevant with regard to state and local public sector pension plans. That clause provides that no state "shall deprive any person of life, liberty, or property without due process of law"1.2 Its applicability to public plans arises generally in two contexts: (1) attempts by a state or local government to reduce the value of the entire accrued pension plan package of a participant, and (2) denials of appliations for benefits, suspensions of benefits, and so on, by the plan without affording minimal due process procedural guarantees to the affected plan participant.
As a preliminary note, many cases which on their facts could involve federal due process considerations are resolved in favor of the plan participant on state constitutional or statutory grounds which relate specifically to the retirement systems of the state, thereby obviating the need for treatment of the due process aspects of the case. For example, Sgaglione, v. Levitt,3 discussed in detail infra, involves a New York statute which purported to impair the value of plan participants' benefits by removing the plan trustee's discretion regarding the investment of a portion of the plan assets. As is apparent, an argument could be made based on these facts that the state took something of value from the plan participants without compensating them for that taking-a prima facie due process violation. Because New York ha's a constitutional provision which explicitly provides that benefits in any state or local retirement system shall not be diminished or impaired,4
1U.S. CONST. amend. XIV.
2 Id.
3 37 N.Y. 2d 507, 337 N.E. 2d 592 (1975).
4 N.Y. CONST. art. V, 7.
(7)






8

the New York court resolves the issue in Syaglione in favor of the plan participants without reaching any due process considerations.
Another example is found in Bakenhus v. City of Seattle.5 Bakenhus, was a police officer whose service commenced in 1925 when Seattle's police pension plan provided for a benefit of 50% of the police officer's salary during his last year of active service. In 1937, the Washington legislature. purported to amend the state retirement laws to provide for a maximum benefit of $125 per month. Upon his retirement in 1950, Bakenhus discovered that the 1937 statutory enactment represented a reduction in his pension benefit of $60 per, month. Despite the apparent due process issues involv 'ed, the Washington Supreme Court found for Bakenhus on a combination of common law contractural and equitable theories of recovery, never finding it necessary to address the due process constitutional issues.
Before an aggrieved public plan participant can make an argument that he has experienced a denial of due process, he must successfully characterize his interest in his plan as a "property" interest.
Although there is clearly a trend towards the characterization of public pension plan interests as "contractural") or "property" rights, there remains a lingering notion that plan interests are more like a gratuity from the employer than an enforceable obligation to the participant. TIhe only Supreme Court case on point, Pennie v. Reis, 6 suggests that until every condition precedent to the actual payment of benefits has been satisfied, the participant's interest in the f und remains a "mere expectancy, created. by the law, and liable to be revoked or destroyed by the same authority." I' As late as 1954, the Pennsylvania Supreme Court noted that "there still lingers a remnant of the ancient idea that a [public] pension is a manifestation of sovereign generosityy" In ('reps v. Board of Firemen's Relief and Retirement Fund Tru stees of Amarillo 9 the Texas Court of Civil Appeals in 1970 approvingly restated 10 an earlier holding of the Texas Supreme Court that "the rule that the right of a pensioner to receive monthly payments from the pension fund . is subordinate to the right of the legislature to abolish the pension system, or diminish the accrued benefits of pensioners thereunder, is undoubtedly the sound rule to be adopted."" The Arkansas Supreme Court in 1973 noted that "technically, a pension constitutes a 'mere gratuity' subject to modification or repeal as opposed to a vested right not subject to such impairment." 12
Despite this authority, the trend is clearly towards viewing interests in governmental plans as property or contractual interests, thereby making them subject to the due process limitations of the 14th Amendment. In Hiickcey v. Pension Board,"3 the Pennsylvania Supreme Court noted that "the pension of today is not a grant of the Republic nor in this case is it a gift of the City Fathers. It is the product of mutual promises between the pensioning authority and the pensioner. . # 14 and "The Legislature may strenghten the
6 4R ~,h 2d G91), 296 P. 2d 36 (1956).
1 1:32 .4 .
7 1, t41.
8 fit-key v. Pensiion 'Board of City of Pittsburgh, 3;78 Pa. 300o, 304, 106 A. 2d 233, 235 (1954).
',W 2d1 4341 (ex. Civ. App. 1970).
wId. at r17.
'("il y of IaNLas v. TrarnMT1.1 129 Toy. 1-70, 158, 101 S.W. 2d 1009, 1013 (1937).
lmw!IIP v. ('hoey, 253 Ark. 926, 09.)0, 18'j S.W. 2d1 7841, 788 (1973).
'0 I)Tpa, jIotO S.
1Supru, note 8 at 305s, 106 A. 2d1 at 235.





9

actuarial fibers but it cannot break the bonds of contractural obligations. The permissible changes, amendments, and alterations provided for by the Legislature can apply only to conditions in the future, and never to the past." 1-1 The Washington Supreme Court in Bakenhuts v. City of Seattle states: "In this state, a pension granted to a public employee is not a gratuity but is deferred compensation for services rendered. ", and refers to the city's obligation to
pay the pension as being of a "contractual nature." 16
Despite an acknowledgement that a public pension constitutes a mere gratuity, the Arkansas Supreme Court in 1973 states: "The retirement pay received by a retired employee . under a system based on voluntary contributions of that employee, represents delayed compensation for services rendered in the past due under a contractural obligation inuring to his benefit, and is not a gratuitous allowance in which the pensioner has no vested right." 11 The New Jersey Supreme Court in 1964 states: "We have no doubt that [public] pension benefits are not a gratuity. . ."1 18 and that "We think there is no profit in dealing in labels such as 'gratuity', 'compensation', 'contract', and 'vested rights'. None fits precisely, and it would be a mistake to choose one and be driven by that choice to some inevit 'able consequence." 19 'The New Jersey Court does go
on, however, to state that "the employee has, a property interest in an existing fund which the State could not simply confiscate."2 In Opinion of the Justices,' the Massachusetts Supreme Court is faced with a pr-oposed state legislative revision of the state's retirement plan in which compulsory contributions by government employees would be raised from 5 percent to 7 percent without any increase in the benefits to be paid. The Massachusetts court first looks to its state constitutional provisions 22 and concludes that the participant's interest in the plan involves a "contractural relationship" and not a "'gratuity"? 3The court goes on to suggest that the interest is not a contract under an "analysis according to Professor Williston's canons"11, but rather "is best understood as meaning that the retirement scheme has generated material expectations on the part of employees and those. expectations should in substance be respected."1 24 Numerous other cases also demonstrate that under the modern, and increasingly followed, view, interests in governmental benefit plans are characterized as propertyt" or "'contractural"), rather than "(gratuitous" in nature.' '
As suggested above'26 the tendency of courts to resolve cases involving due process issues on specific state constitutional or statutory provisions relating directly to benefit plans rather than on the more general due process clause of the Fourteenth Amendment has
16 Supra, note 8 at 309, 106 A. 2d at 237.
Is Supra, note 5 at 698, 296 P. 2d at 538.
17 Su~pra, note 12 at 930, 489 S.W. 2d at 787--88.
is Spina v. Consolidated Police and Firemen's Pension Fund 'Commission, 41 N.J. 391, 402, 197 A. 2d 169, 175 (1964).
19 Id. at 401, 197 A. 2d at 174.
20 Id. at 402, 197 A. 2d at 175.
21 364 Mass. 847, 303 N. E. 2d 3290 (Mass. 1973).
22 Mass. Gen. Law Ann., ch. 32, 25(5).
23 Supra, note 21 at 860, 303 N.E. 2d at 327.
24 Supra, note 21 at 861, 303 N.E. 2d at 327-28.
25 See e.g. YeaZell V. Copins, 98 Ariz. 109,402 P. 2d 541 (196); Kern v. City of Long Beach, 29 Cal. 2d 848, 179 P. 2d 799 (1947).
26 Supra, p. 2.





10

resulted,in a very limited number of cases in which there has been a resolution of the due process issues. Whether a plan interest is "property" for due process purposes and what level of procedural due process must be given participants in various situations (e.g. denial of claims for benefits) are largely unresolved issues. The few decisions that discuss these issues do provide some guidance nonetheless.
Siletti v. New York City Employees' Retirement System 27 addresses both of the due process issues just described. Plaintiff's application for accident disability benefits was denied without plaintiff's having had an opportunity to present evidence under oath, to have an evidentiary hearing, and so on. The District Court first considered whether plaintiff's interest in the disability plan was sufficiently a property interest to merit due process protection. It concluded that "the disability benefits at issue are entitlements created by statute for the benefit of persons meeting the specified qualifications. The Fourteenth Amendment's protection of property has been broadly read to expand protection to such entitlements." 28 The Court goes on to consider what procedural safeguards Siletti must be afforded, and concludes that a full adversary or trial type hearing is not required because questions of credibility and veracity of witnesses are not likely to be involved, and hence plaintiff has little to gain from crossexamining witnesses, taking testimony under oath, and so on. Because the defendant's denial of plaintiff's application apparently occurred in a fair proceeding, the Court refuses to allow Siletti any of the proceedings he sought.193
In Loweher v. Beame,0 plaintiff sought to be awarded serviceconnected disability status. Defendant's Medical Board twice examined plaintiff, each time denying her request to be represented by an attorney and to present witnesses, but allowing her to submit her own medical records. Reports of a third examination, done by a physician to whom plaintiff was referred by the Medical Board, were denied to plaintiff.The court never considers whether the nature of plaintiff's interest is property", but apparently assumes as much, for the court proceeds directly to consider whether plaintiff was given satisfactory procedural safeguards. Contrary to the conclusion in Siletti, the court here finds that "Although not entitled to a full adversarial hearing, plaintiff was entitled to know the evidence before the Medical Board. The right to be advised of the evidence upon which the Retirement System makes its determination is implicit in procedures which afford due process of law." 1 3
The New York Court of Appeals has twice, albeit briefly, addressed these issues in recent years. In Mesehino v. Lowery y32 the Court of Appeals finds that an interest in a disability plan does require that procedural safeguards be present, but it finds that the opportunity to present documentary and oral testimony was sufficient.
In Balash v. New York City Employee's Retirement System,33 how27 401 F. Supp. 162 (S.D.N.Y. 1975).
'A Id. at 167 (emf11pha~is addled).
8405 F. Supp. 1247 (S.D.N.Y. 1975).
"Id. at 1241).
'2it N.Y. 2:1 772 2(4) N.E. 2,! 8.- (N.Y. 1972).
U34 N.Y. 216c51, 311, N.E. 21 C,9 (1971).





11

ever, the Court of Appeals appears to take a somewhat broader view of what degree of procedural safeguards a plan participant must be presented before his "property" (i.e. interest in benefit plan) is quantified. Petitioner, a disability plan participant, was retired from his active employee status and placed on ordinary disability, but given no written notice of any reason for the proposed retirement and no opportunity to present medicalor other evidence to either the medical panel or its parent Board of Trustees of the Retirement System. Nor was he, presented with any medical reports regarding that condition which led the medical panel to itsconclusions regarding his medical condition. Similarly, after his involuntary retirement, he was not given any written statements of the reasons for the action. The Court of Appeals, in finding for petitioner, states:
Petitioner should have been advised of the charges against him and the evidence on which they were based, afforded meaningful opportunity to present documentary or other evidence in his favor at least to the Board of Trustees, and advised of the reasons for their determination ... Due process does not require that petitioner has been afforded a full-blown adversary hearing with the right to cross-examine the Dsychiatrists whose reports formed the basis of the medical papers certification and the board's subsequent determination. He had, however, the right to be informed of the substance of those reports and should have been given an opportunity, at least before the Board of Trustees, to controvert the conclusions they contained
Despite the limited number of cases discussing the due process issue, a number of tentative conclusions may be drawn. It appears that a participant's interest in his state or local governmental plan is sufficiently an interest 'in "property" to trigger the safeguards against deprivations contained in the Fourteenth Amendment. It also
-appears that in most proceedings in which the nature of a participant's benefit plan interest is adjudicated, the participant must be given notice of the nature of the proceeding and the evidence upon ch
the gOVeMM Dt expects to reach its conclusion, and afforded an opportunity, not necessarily in person, to present evidence to the officials adjudicating the participant's claim. A full adversarial proceedmig *in 'Most instances,, it appears, will not be required.
That interests *in governmental p nsion and welfare plans axe receiving this treatment is not surprising, in light of recent Supreme Court decisions which have taken a more and more expansive view of what kinds of interests represent "liberty" and "property" for due process purposes.35 Further refinement of the requirements involving the degree of procedural due process that must be provided can be e ec ed as courts axe presented with cases involving indigent plan participants, illiterate plan 'participants, adjudications of complex benefit formulas, selection of complicated retirement options, and so on.
A closely related issue, involving both the due process clause of the Fourteenth Amendment as well as the non-impairment of contractual obligations clause of Article 1,311 is to what degree a state
34Id. at 656, N.E. 2d at 649.
2-5 See, e.g., Goldberg v. Kelly, 397 U.S. 254 (1970).
34 U.S. CONST. art. I, 10, cl. 1: "No State shall enter into any Treaty, Alliance, or Confederation; grant Letters of Marque and Reprisal; coin Money; emit Bills of Credit; make anything but gold and silver Coin a Tender in Payment of Debts; pass any Bill of Attainder, ex post facto Law, or Law impairing the Obligation of Contracts, or grant any Title of Nobility."
74-365-78-2






12

may alter its benefit plan without thereby unconstitutionally either taking property of plan participants or impairing its contractural obligations to plan participants. As might be expected, the issue has Yet to be fully resolved. But two state supreme courts in recent years have addressed it in a manner that appears sound and likely to be followed in future years.
In Allen v. City of Long Beach,7 the California Supreme Court was presented with a charter amendment of Long Beach City which substantially and retroactively reduced the benefit and other pension entitlements of participants. Plan participants attacked the charter amendment on the ground that it unconstitutionally impaired the contractural rights, in the form of interests in the pension plan, which j)articipa-nts had attained.
The California Supreme Court agrees that the charter amendment represented an unconstitutional impairment, and states:
An employee's vested contractual pension rights may be modified prior to retirement for the purpose of keeping a pension system flexible to permit adjustments in accord with changing conditions and at the same time maintain the integrity of the system... Such modifications must be reasonable, and it is for the courts to determine upon the facts of each case what constitutes a permissible change. To be sustained as reasonable, alterations of employees' pension rights must bear some material relation to the theory of a pension system and its successful operation, and changes in a pension plan which result in disadvantage to employees should be accompanied by comparable new advantages . In the present case it appears that . [the amendment] substantially decreases plaintiffs' pension rights without offering any commensurate advantages, and there is no evidence or claim that the changes enacted bear any material relation to the integrity or successful operation of the pension system... .1
This statement that modifications in the plan are permitted so long as the net value of the participant's interest remains essentially unchanged is restated by the Massachusetts Supreme Court in Opinion of the Justices." The Massachusetts House of Representatives requested an advisory opinion on a proposed increase in the participants' rate of contribution (from 5% of salary to 7% of salary) with no increase in benefit or other plan provisions. The Massachusetts Court considers the proposed change in light of the impairment of contract clause, or, "what amounts to much the same thing," 10 the due process clause of the Federal and state Constitutions, and concludes:
Legislation which would materially increase present members' contributions without any increase of the allowances finally payable to those members or any other adjustments carrying advantages to them, appears to be presumptively invalid-invalid, that is to say, unless saved by the reserved police powers... That the maintenance of a retirement plan is heavily burdening a governmental unit has not itself been permitted to serve as justification for a scaling down of benefits figuring in the 'contract,' although no case presenting proof of a catastrophic condition of the public finances has been put... .41
Thus it appears that no significant depreciation in the overall value of a benefit package is permitted in those plans in which interests are characterized as contractual in nature, with regard to past accruals of present and retired participants and also with regard to future accruals
45 Cal. 2d 12,, 2-7 1'. 211 765 (195.).
4 Id. 131, 287 1'. 2d at 77.
Supra, note 21.
40 Siupr, note 21 at si3, "03 N.E. 2d at 3 2 .
41 Suipra, note 21 at 1, 3(}3 N.E. 2d at 32-.'.






13

of pre-sent plan participants. Of interest and potential importance are the suggestions in both the California and Massachusetts opinions that impairments of contractual rights might be permitted upon the occurrence of some "catastrophic conditions," 42 or if necessary "to the functioning and integrity of the pension system." 41Whether the kind of fiscal difficulties experienced by numerous muncipalities in recent years would approach the thresholds implied in these opinions is problematic. In any event, it is clear that the due process clause of the Fourteenth Amendment and the impairment of contracts clause of Article I will remain of significance to public employee retirement systems.
FOURTEENTH AMENDMENT-EQUAL PROTECTION
Another constitutional issue involves mandatory retirement as an element of pension plans sponsored by state and local government employers, and the Equal Protection Clause of the Fourteenth Amendment.' The Equal Protection Clause requires that where a fundamental right or a suspect classification is involved, the government must not discriminate between classes of citizens unless there is a compelling public interest in doing so; i.e., the classification must withstand strict scrutiny.2 If neither a fundamental right nor a suspect classification is involved, then a government, consonant with equal protection requirements, may 'discriminate between classes of citizens if such discrimination is predicated on a rational basis, rather than justified by a compelling public interest.'
The Supreme Court in Massachusetts Board of Retirement v. Murgia4
squarely addresses the equal protection issue in the context of a state government's mandatory retirement policy. Massachusetts law required that all state police officers must be retired upon attaining age 50--1 Murgia, a state police officer who was retired upon his 50th birthday, alleged that this mandatory retirement policy denied him equal protection of the law. The Supreme Court first decides that "rationality is the proper standard by which to test whether compulsory retirement at age 50 violates equal protection", and then
finds the Massachusetts classification meets this rationality standard.* The Court goes on to state:
[Sitrict scrutiny is not the proper test for determining whether the mandatory retirement provision denies appellee equal protection. San A ntonio Independent School District v. Rodriguez, 411 U.S. 1, 16 (1973), reaffirmed that equal protection analysis requires strict scrutiny of a legislative classification only when the classification impermissibly interferes with a fundamental right or operates to the peculiar disadvantage of a suspect class. Mandatory retirement at age 50 under the Massachusetts statute involves neither situation.
This court's decisions give no support to the proposition that a right of governmental employment per se is fundamental . Accordingly, we have expressly stated that a standard less than strict scrutiny "has consistently been applied to state legislation restricting the availability of employment opportunities"
0 1d.
43 Sutpra, note 37 at 133, 287 P. 2d at 788.
I U.S. CONST. amend. XIV.
2 San Antonio Independent School Dist. v. Rodriguez, 411 U.S. 1 (1973).
3 Dandridge v. Williams, 397 U.S. 471 (1970).
4 427 U.S. 307 (1976).
5 Mass. Gen. Laws. chi. 32 26(3) (a) (1966).
6 Supra, note 4 at 312.





14
Nor does the class of uniformed state police officers over 50 constitute a suspect class for purposes of equal protection analysis. Rodriguez observed that a suspect class is one "saddled with such disabilities, or subjected to such a history of purposeful unequal treatment, or regulated to such a position of political powerlessness as to command extraordinary protection from the majoritarian political process." While the treatment of the aged in this Nation has not been wholly free of discrimination, such persons, unlike, say, those who have been discriminated against on the basis of race or national origin, have not experienced a history of purposeful unequal treatment or been subjected to unique disabilities on the basis of stereotyped characteristics not truly indicative of their abilities. The class subject to the compulsory retirement feature of the Massachusetts statute consists of uniformed state police officers over the age of 50. It cannot be said to discriminate only against the elderly. Rather, it draws the line at a certain age in middle life. But even old age does not define a "discrete and insular" group, in need of "extraordinary protection from the majoritarian political process." Instead, it marks a stage that each of us will reach if we live out our normal span. Even if the statute could be said to impose a penalty upon a class defined as the aged, it would not impose a distinction sufficiently akin to those classifications that we have found suspect to call for strict judicial scrutiny.7
The Court the 'n goes on to demonstrate how the Massachusetts
classification is a rational one."
Murgia, then, clearly affirms that age discrimination by a government, at least in the form of mandatory retirement, and probably in any form, is to be tested for equal protection purposes against a simple "rational" standard, rather than against a more trying "strict scrutiny" or "compelling public interest" standard.
The Court has not, in cases following Murgia, withdrawn from Murgia in any way. In conclusion, then, it may be said that federal constitutional law does not significantly affect whether state and local government retirement systems can discriminate on the basis of age.9
An equal protection issue is also present in the growing number of cases involving governmental pension and welfare plans in which it is alleged that a governmental plan's treatment of women differently from men represents a differentiation between classes of citizensmen and women-that is prohibited by the mandate of the 14th Amendment that no state deny its citizens the equal protection of the law.
The Supreme Court has addressed this precise issue in recent years. In Geduldig v. Aiello,'0 the Court considers a disability insurance program that excluded from coverage disabilities due to pregnancy as well as a number of other medical conditions. It was contended that the exclusion of pregnancy related disabilities violated the Equal Protection Clause. In directly rejecting this contention, the Supreme Court states that the exclusion of pregnancy-related disabilities from the set of risks insured by California in its plan does not represent the kind of invidious discrimination prohibited by the Equal Protection Clause. So long as the disability program is "rationally supportable", the Court holds, no equal protection violation exists."1
7 Su pra, note 4 at 312-14. Footnotes omitted.
I Supra, note 4 at 315.
Buta cf. I,,ad ley v. Vance, Civil Action No. 76-4035 (D.C. Cir. June 23,' 1977), in which it i s held that mandatory retiroent at age (JO for foreign service personnel is not National and hence violates the Equal Protection Clause.
11 Id. at 495.






15

The Court again addresses the issue of whether sex related distinctions in plans represent discrimination on the basis of gender, albeit not in irn Equal Protection context, in General Electric Co. v. Gilbert,'12 also discussed elsewhere in this report." The Court, first notes that Geduldig v. Aiello deals with a disability plan very similar to the General Electric plan, and indicates that therefore the analysis the Court made in Geduldig, to the effect that the exclusion of pregnancy from a disability plan that otherwise did not discriminate against a definable class or gru terms of the overall risk protection afforded by the plan, wZould be "~quite relevant" in considering the General Electric plan.' As it held in Geduldig, the Supreme Court in Gilbert finds the General Electric plan is not a pretext "designed to effect an invidious discrimination against the members of one sex or another."'5 Nor is the distinction drawn in the G.E. plan, the Court adds, a sex-based distinction which is neutral on its face but in reality a subterfuge to accomplish a forbidden discrimi nation.
In terms of sex discrimination and the Equal Protection Clause generally, the Court has not indicated with any clarity whether sex discrimination by a governmental entity must be justified by a "compelling state interest" and subject to "strict scrutiny" or merely reflect a "rational relationship to a legitimate governmental interest." 18 The Court appears to be headed towards a new, middle standard in which sex based discrimination must serve "important governmental objectives" and be "substantially related" to the achievement of those objectives.'7
The Equal Protection issue is obviously closely related to the Title VII, sex discrimination, issue. Whether a state pension plan could discriminate on the basis of sex in a manner which violates the Title VII prohibitions but not the Equal Protection Clause is not clear. Presumably Congress, in making state and local governments subject to Title VII, intended to prohibit certain employer practices that were not already prohibited by the Equal Protection Clause.'8
The Supreme Court has agreed to review '9 City of Los Angeles v. Manhart, discussed at length in this report's discussion of Title VII. However, Manhart does not involve the Equal Protection Clause. Plaintiff in that case has challenged the defendant's pension plan practice only on the basis of Title VIL.20 Thus it appears unlikely that the Supreme Court in Manhart will clarify the relationship between Title VII and the Equal Protection Clause in the context of sex discrimination in a governmental employer's pension or welfare benefit plans.
In declining to review Reilly v. Robertson '2' the Supreme Court22 over12 429 U.S. 12 (1976).
Is Infra, P. 24.
14 Supra, note 12 at 133.
isSupra note 12 at 134, quoting Geduldig v. Aiello.
is &ee, e.g., Frontiero v. Richardson, 411 U.S. 677 (1973), and Reed v. Reed, 404 U.S. 71 (1971). 17' Califano, v. Goldfarb, 430 U.S. 199 (1977); Califano v. Webster, 430 U.S. 313 (1977); Craig v. Boren, 429 U.S. 190 (1976).
"iWashington v. Davis, 426 U.S. 229 (1976).
10 U.S.L.W. 3214 (1977) (No. 76-1810).
20 Infra, p. 23.
21 56 Ind. Dec. 400, 360 N.E. 2d 171 (1977).
22 46 U.S.L.W. 3215 (1977).






16

looked an opportunity to address the Equal Protection Clause issue absent in Manhart. Reilly involves the use of sex-based mortality tables by the Indiana State Teachers' Retirement Fund and the differing benefits received under the plan by similarly, situated men and women. In a suit brought by a woman annuitant in the plan, the Indiana Supreme Court first concludes that for purposes of the federal Equal Protection Clause the "fair and substantial relation" standard is to be applied to the plan's use of sex-based mortality tables, and that a "substantial distinction" and "manifestly unjust or unreasonable" standard is to be used for purposes of the Indiana Constitution.2 It then proceeds to affirm the trial court's conclusion "that it was arbitrary and without rational basis for the fund to classify annuitants by sex . ."2 and hence violative of both the federal and state Equal Protection clauses. It is possible that the U.S. Supreme Court refused to review the Indiana Supreme Court decision because the Indiana equal -protection provisions would control the result even if the Indiana Court's treatment of the federal Equal Protection Clause were erroneous.
In conclusion, the significance of the Equal Protection Clause of the 14th Amendment in the context of governmental plans and sex discrimination is far from clear. It does appear that the Equal Protection Clause permits a state or local government to discriminate on the basis of gender only when such discrimination serves important governmental objectives and is substantially related to the achievement of those objectives. The meaning of gender-based discrimination in pension and welfare plans, however, remains very uncertain, as does the resolution of the issue of whether the sex based discrimination prohibited by Title VII is identical to the discrimination prohibited by the Equal Protection Clause.

ELEVENTH AMENDMENT
The Eleventh Amendment to the Constitution'I at present affects public employee retirement systems to a slight extent. Enactment of federal legislation regulating state and local governmental pension plans would heighten its importance with regard to public plans.
The Eleventh Amendment until recently has been interpreted
generally to bar a United States District Court from awarding monetary damages to be paid by the State .2 Suits in federal court against state officials to obtain prospective relief are not prohibited.3 The law is also clear that the aJphlcability of the 11th Amendment immunity from suit turns on whether the state is the real party in interest, even though it may not be named directly in the Suit itself,4 and that an award of money damages cannot be disguised as some sort of equitable restitution for the purpose of evading -the jurisdictional limitations of the 11th Amendment.' Thus, until recently,
23 S upra, note 21 at 4051 O0 N.E. 2d at 174-75.
Sa SPra. note 21 at 407. :360 N. E, 2d at 176.
U-S. CONST. amenid. XI: ..T1he Judicial power of the United States shall not be construed to exten
to aiiy suit in law or equity, COnM~etwed or prosecuted against one of the United States by Citizens of aniot 1wr State, or hy (Cit izon or Subhjects of any Foreign State."
2 1.o(rd MNotor (Co. V. 1)ept of Treasunry of Indt~iana, 323 U.S. 4159 (19-13). SEx part, Youngf, 209) (.S. 12:3 (1908).
4 S pra, ioteo 2.
~ Ldhnanv. Jordan, 415 U.S. G-51 (1974).






17

participants, assuming they satisfy other jurisdictional requirements, could sue their state plan officials in federal court for breaches of their duties under state law, and receive injunctive and other prospective relief, but not money damages. Local governments do not enjoy the benefits of the Eleventh Amendment."
In Ktzpatrick V. Bitzer, 7 the Supreme Court is faced with an award of money damages (attorney's fees) to be paid by the state retirement fund as part of the judgment that plaintiff, a plan participant, received in his successful Title VII sex discrimination suit against the state pension system.
The Court holds that the immune y from an award of money damages which the state enjoys under the Eleventh Amendment is limited by the enforcement authorization in the Fourteenth Amendment, and hence the award of legal fees as part of a judgement based on Title VII was within the jurisdiction of the district court."
This holding is of considerable significance for purposes of any of the various formulations of a PERISA. By enacting federal legislation regulating public employee retirement systems pursuant to the Fourteenth Amendment, Congress can remove the imnitinity of states from money damage awards by federal courts, thereby enabling the retrospective aspects (e.g. compensation for fiduciary breach) of any federal legislation to be litigated and enforced in federal court. This result is highly desirable in that it permits the developm.--.nt of a more unified and consistent interpretation of national law. It also avoids the situation whereby an aggrieved plan participant might be forced to turn to federal court for injunctive relief and then to state .court to receive money damages.
An interesting observation is made by Justice Stevens in his concurrence, in which he states that the attenuated relationship between the payment of attorney's fees by the state pension fund and the state's funding obligation makes the Eleventh Amendment "defense" in this case inapplicable, since the award would not be paid by the state in reality, but by the plan." Whether the Court will in the future focus more closely on the relationship between the plan and the state government is unclear.

BASES FEDERAL JURISDICTION
As the prospect of federal legislation regulating public employee retirement systems becomes more likely, it is necessary to carefully consider those bases of jurisdiction for such regulation available to the Congress.
The most obvious jurisdictional basis for Congressional regulation of public employee retirement systems is the Comm rce Clause power f ound in Axticle L section 8, clause 3 of the Constitution. Congress, by that clause, is generally given the power to regulate interstate comm ce. Over the years the meaning of interstate commerce has expanded significantly. In a case involving provisions of the Civil Rights Act of 1964,1 the Supreme Court summarized the meaning of "interstate comm for purposes of the Commerce Clause: "In
Gilliam v. City of Omaha, 524 F. 2d 1013 (8th Cir. 1975); Fay v. Fitzgerald, 478 F. 2d 181 (2d Cir. 1973).
7 427 U.S. 445 (1976).
8 Id. at 456.
2 Id. at 459-60.
142 U.S.C. 2000a et seq. (1970).






18

short, the determinative test of the exercise of power by the Congress under the Commerce Clause is simply whether the activity sought to be regulated is 'commerce which concerns more states than one' and has a real and substantial relation to the national interest."2 Given the tremendous impact public employee retirement systems have on the securities markets and the national economy, the frequent movement between various states of public plan participants, and numerous other factors all demonstrating that public employee retirement systems involve more than the state in which the plan is located, it is clear that public plans involve interstate commerce, and hence may be regulated by Congress under its Commerce Clause power.
Recent court decisions, most prominently National League of Cities v. Usery,3 make clear, however, that Congress' Commerce Clause power is not unlimited.
At issue in National League of Cities were 1974 amendments to the Fair Labor Standards Act.- These amendments purported to make state and local governments subject to the minimum wage and overtime provisions of the Fair Labor Standards Act, thereby directly and significantly affecting the fiscal operations of state and local governments. The National League of Cities, joined by numerous state and local governments, argued that the Congressional action under the Commerce Clause power was unconstitutional because it improperly intruded into basic state government functions, a power not given the Congress but rather reserved to the states in the 10th Amendment.6
Despite its earlier affirmation of 1) Congressional extension of minimum wage and overtime provisions to state hospitals, schools, and institutions Jand 2) Congressional limitation of state wage and salary increases,8 and its unamended pronouncement many years earlier in United States v. California 9 that "there is no such limitation upon the plenary power [of Congress] to regulate commerce",10 the Supreme Court agrees with appellants' contention and finds the Congressional exercise unconstitutional.
The Court appears to focus on the relationship between the 10th Amendment reservation to the states and the grant of power to the Federal Government in the Commerce Clause. The majority opinion first notes with approval the suggestion by Justice Marshall in note 7 in Fry v. United States that "The [10th] Amendment expressly declares the constitutional policy that Congress may not exercise power in a fashion that impairs the States' integrity or their ability to function effectively in a federal system." 1l The Court then proceeds to focus on the effect the FLSA amendments would have on the states' ability to function under the burden of the minimum wage
and overtime provisions, stating that such extension "will impermissibly interfere with the integral governmental functions" 1 of states and their political subdivisions.
2 Heart of Atlanta Motel, Inc., 379 U.S. 241, 255 (1961).
9 426 U.S. 833 (1976).
4 Pub. L. No. 93-259, 88 Stat. 55 (1974).
6 29 U.S.C. 201 et seq. (1970).
6 U.S. CONST. amend X; "The powers not delegated to the United States by the Consitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people." Maryland v. Wirtz, 392 U.S. 183 (1968).
Fry v. United States, 421 U.S. 512 (1975).
2 297 U.S. 175 (1936).
2 I ld at 1 5.
11 Supra, te 8
12 S'upra, note 3 at 851.





19

In considering the ,FLSA amendments in light of the clarified
relationship between the 10th Amendment and the Commerce Clause, the Court states: "This exercise of congressional authority does not comport~ with the federal system of government embodied in the Constitution. We hold that insofar as the challenged amendments operate to directly displace the States' freedom to structure integral operations in areas of traditional government functions, they are not within the authority granted Congress by Art. I, 8, cl. 3V' 13
The Court, in short, holds that the tension between Congressional power under the Commerce Clause and the reservation of power to the states in the 10th Amendment must be resolved in favor of the states when the Congressional exercise threatens the ability of the states to function as sovereigns in the federal relationship.
A number of other aspects of National League of Cities should be noted. The decision is an extremely close one. Only four Justices join with Justice Rehnquist in the majority. opinion, and one of them, Justice Blackmun, in a concurring opinion expresses strong reservations over portions of the Court's decision.
Also, courts considering 10th Amendment arguments have generally given National League of Cities an extremely narrow interpretation, refusing to find that various Congressional enactments such as the Equal Pay Act 14 and the Age Discrimination in Employment Act 15 impose sufficient burdens on state governments to threaten the state and local governments' ability to function, the finding required for Congressional actions based on the Commerce Clause to exceed the limitation placed on that power by the 10th Amendment. 16
With regard to Congressional enactment of a Public Employee Retirement Income Security Act (PERISA) based on Commerce
Clasejurisdiction, it appears that only a full, immediate funding requirement would even begin to affect the fiscal or other operations of stateand local governments in a manner which might threaten the viability of a state to continue as a sovereign. Clearly federal legislation'limited to*such items as reporting, disclosure, and fiduciary responsibility would produce a very slight cost impact in terms of compliance by state and local governments. In fact, a net reduction in cost might be achieved as existing asset management and investment techniques were replaced with federally mandated practices that produced greater income, more efficient management techniques, and so on. Even federal vesting requirements, in the absence of strict funding standards, would probably not reach the level of intrusion in basic state functions which the Supreme Court found in League of Cities. Further, it is conceivable that legislation mandating a relatively long-term funding requirement (e.g. forty years to fund past service liabilities) would be permissible under Commerce Clause jurisdiction. As will be discussed, infra, given the ability of Congress to act pursuant
13 Supra, note 3 at 852.
14 29 U.S. C. 206 (d) (1970).
IS 29 U. ,S.C. 621 et seq. (1970).
Is -Se, e.g., Usery v. Bd. of Educ. of Salt Lake City, 421 F. Supp. 718 (D. Utah 1976); Usery v. Bettendorf Community School Dist., 423 F. Supp. 637 (S.D. Iowa 1976); Christensen v. Iowa, 417 F. Supp. 423 (N.D. Iowa 1976); Usery v. Allegheny County Inst. Dist. 544 F. 2d 148 (3d Cir. 1976), cert. denied, 45 U.S.L.W. 3651 (1977); Usery v. Charlestown County School Sys., No. 76-2340 (4th Cir. July 25, 1977); Usery v. Fort Madison School Dist., No. C 75-61-1 (S.D. Iowa, motion to dismiss denied Sept. 1, 1976); Usery v. Sioux City Cornmullnity School Dist., No. C 76-4024 (N.D. Iowa, motion to dismiss denied Aug. 20, 1976); Riley v. Univ. of Lowell, Civ. No. 76-2118-M (D. Mass., motion to dismiss denied July 22, 1976).





20

to its power under the 14th Amendment, the issue of limited Commerce Clause power, for purposes of federal regulation of state and local pension systems, is largely academic.
In conclusion, it appears that Congressional power to enact legislation, including some sort of a "PERISA", pursuant to the Commerce Clause is limited by the 10th Amendment only when the Congressional enactment so vitally affects a basic state or local government function that the capacity of the state to function as a sovereign in the federal relationship is severely threatened.
A second source of federal jurisdiction, admittedly cumbersome and indirect, by which the federal government can constitutionally regulate public employee retirement systems, involves the federal taxing power.
Congress clearly has the authority to tax income.'7 It is equally clear that Congress may grant certain deductions and exemptions from such income tax, s and that "Whether and to what extent deductions shall be allowed depends upon legislative grace; and only as there is clear provision therefore can any particular deduction be allowed." 19
Congress could currently impute to participants in governmental plans the value of their vested and funded interests in such plans, as well as the income generated by such interests. It could then condition a deferral, deduction, or exemption to currently recognizing that income upon the satisfaction by the plan of various conditions. Such a methodology would obviously generate tremendous pressure by plan participants on public plans to comply with whatever federal conditions must be met to gain favorable tax treatment for the participants.
The disadvantages of this kind of regulation are apparent. The regulation is not of the plan directly; it relies on the desire of plan participants to receive favorable tax treatment to force plan sponsors and administrators to satisfy whatever conditions must be satisfied in order to receive beneficial tax treatment. It is cumbersome in that the federal government does not mandate that governmental plans must meet specified standards, but rather relies on a permissive, conditional, tax exemption or deferral over which it has no direct control to achieve the dlesired standards in governmental plans. The government's primary concern would be to enforce the tax law; achievement of socially desirable plan operations and provisions would be a secondary concern. Finally, plan participants would be excluded from the regulatory process in that the relationship, as in most tax matters, would be exclusively between the taxpayer (the plan) and the Internal Revenue Service.
These dsadvantages notwithstanding, the taxing power represents a basis for federal regulation of public employee retirement systems which completely avoids any limitations on Congressional jurisdiction imposed by the 10th Amiendment.
A third, and extremely effective, source of power through which Congress can fully regulate public employee retirement systems is found in the Fourteenth Amendment of the Constitution. That amendment states:

1 I ... ONST. aCw0Nd XVI.
b- Ilekin V. 1)isirivI of ( olmnhia, 461 F. 2(l 1615 (D.C. Cir. 1972). Now Colonial Ice (o. v. Itclvcring, 692 U.S. 435, 440 (1934).





21
Section 1. All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States and of the State wherein they reside. No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States; nor shall any State deprive any person of life, liberty, or property, without due process of law; nor deny to any person within its jurisdiction the equal protection of the laws ....
Section 5. The Congress shall have power to enforce, by appropriate legislation, the provisions of this article.20
By finding that the rights to be protected by a Public Employee Retirement Income Security Act are property rights, or rights in the nature of liberty, or that the purpose of the legislation is to afford citizens equal protection of the laws, Congress can enact PERISA legislation pursuant to section 5 of the 14th Amendment rather than pursuant to the Commerce Clause power.
Recent judicial decisions support this methodology as a means of avoiding the limitation placed on Congressional power by National League of Cities.
In South Carolina v. Katzenbach,2' the Court was faced with the argument that the Voting Rights Act of 1965 22 was unconstitutional in that it improperly purported to be enacted pursuant to the power given Congress in the identical enforcement authorization of the 15th Amendment. In considering whether Congressional action is appropriate under the enabling sections of constitutional amendments, the Court states: "As against the reserved powers of the States, Congress may use any rational means to effectuate the constititional prohibition. 2 S" 23
A number of recent Court decisions 24 support the theory that the constitutional prohibitions addressed in section 1 of Amendment XIV which can be addressed by Congress legislatively pursuant to section 5 go well beyond the race oriented injustices prominent at the time the 14th Amendment was enacted and ratified.
Arbitrary treatment of citizens, disparate treatment of various classes of citizens, and the deprivation of rights that are in the nature of property or interests that are in the nature of liberty are all among the Constitutional prohibitions contained in section 1 of the 14th Amendment and among the prohibitions which Congress is empowered to address legislatively in section 5 of that amendment. Clearly any PERISA legislation would directly affect each of these prohibitions.
The repeated willingness of the Supreme Court 25 to apply the prohibitions of section 1 to events, conditions, and processes never contemplated by Congress when it enacted the 14th Amendment in 1868 strongly supports the notion that section 5 gives Congress the power to address those .concerns through legislation. .
Moreover, it is increasingly clear that Congress can affect integral functions of state and local governments pursuant to its power under section 5 of the 14th Amendment even though such legislation may
210 U.S. CONST. amend. XIV.
21 383 U.S. 301 (1966).
2242 U.S.C. 1971, 1973 et seq. (1970).
2 Supra, note 21 at 324.
24 See e.g., Usery v. Board of Education of Salt Lake City, 461 F. Supp. 718 (D. Utah 1976): Usery v. Bettendorf Community School Dist., 423 F. Supp. 637 (S.D. Iowa 1976); Usery v. Allegheny County Inst. Dist., 541 F. 2d 148 (3d Cir. 1976); Usery v. Meyer Memorial llosp., 428 F. Supp. 1368 ( .D.N.Y. 1977).
25 See e.g., Pcrry v. Sindermann; 408 U.S. 593 (1972) in which a strong expectation of continued employment is held to involve a property right, and Goldberg v. Kelly, 397 U.S. 254 (1970), in which it is suggested that welfare entitlements are a form of property.





22t

be beyond the power granted to Congress under the Commerce Clause. National League of Cities v. Usery26 itself explicitly refrains from addressing this issue. But only four days after deciding National Lea guie of Cities, the Supreme Court in Fitzpatrick v. Bitzer 27 (also written by Justice Rehnquist), in recognizing that the 1972 amendments to Title VII of the Civil Rights Act of 1964 are enacted pursuant to section 5 of the 14th Amendment, states: "We think that Congress may, in determining what is 'appropriate legislation' for the purpose of enforcing the provisions of the Fourteenth Amendment, provide for . [measures] which are constititionally impermissible in other contexts." 281 Thus federal legislation that is beyond one federal jurisdictional basis (the Commerce Clause), may nonetheless come within another federal jurisdictional basis (the enforcement sect-ion of the 14th Amendment), to accomplish what was prohibited by the 11 th Amendment .29 There is no reason to believe that in enacting a PERISA Congress could not use section 5 of the 14th Amendment to accomplish what might be prohibited by the 10th Amendment.
If the National League of Cities decision does in some manner limit the kind of regulation of state and local government retirement plans which Congress can legislate, the Fourteenth Amendment's grant of enforcement power provides Congress with an alternative jurisdictional basis that is ample to fully regulate the universe of public employee retirement systems.
A fourth means by which Congress can effectively regulate governmental plans involves the placing of additional conditions onto federal revenue sharing grants made to state and local governments.
Certain conditions relating to the use by local governments of revenue sharing funds have existed in the past, and others continue to exist today.30 For instance, local and state governments receiving funds must be audited independently every three years in accordance with generally accepted accounting principles,"' and must publicize proposed uses of revenue sharing funds.3 Although Congress has been unwilling to attach conditions onto its revenue sharing grants, it clearly could require that retirement systems of state and local governments conform to specified standards as a condition to the payment of revenue sharing funds to the governmental entity. Given the huge sums involved in the revenue sharing program, 33 and the fact that revenue sharing funds are often used to fund state and local government pension programs, attaching conditions to revenue sharing grants represents an effective method by which the federal government could induce state and local government plans to conform their -pension plans to meet certain standards.

CIVIL RIGHTS ACT OF 1964-TITLE VII
Federal law regulating employment practices significantly affects the operation of state and local government pension plans. Section
26 Supra, note 31 at footnote 17.
27 4 27 U. -S. 445 (1976).
2A Id. at 456.
26 Supra, p. 16.
30 Infra, p. 36.
31 Act of Oct. 13, 1976, Pub. L. No. 94-488 (to be codified in 31 U.S.C. 1243(c) (1)). 32 Act of Oct. 1:3. 11)76, Pub. L. No. 941-4SS (to be codified in 31 U.S.C. 1241(c) (1)). 3- Approximately $6 '.9 billion of entitlement payments were authorized under the revenue sharing program for fiscal year 1977. S. Rep. No. 94--1207, 94th Cong., 2d Sess. 1-2, reprinted in 1976 L.S.C.C.A.N.
)51~i-52.






23O

703 (a) of the Civil Rights Act of 19641 'provides that it shall be unlawful for an employer to discriminate against an employee in an employment practice on the basis of race, religion, sex, or national ori gin. The 1964 Act was amended in 1972 2 to bring state and local govern,. ments, as employers, within the scope of employers subject to the Act.
It is undisputed that a government's administration of its retirement system represents an employment practice within the meaning of Title VII.3 Accordingly, it is clear that a-retirement system may not discriminate on the basis of the sex of a plan participant. But what constitutes unlawful discrimination on the basis of sex, in the context of a retirement system, is far from clear.
The-Supreme Court, in Fitzpatrick v. Bitzer,4 recently considered only 11th Amendment issues, and affirmed lower court decisions holding that it is an unlawful employment practice for an employer to maintain a retirement system that allows women to retire earlier than men and that provides higher rate differentials for women who elect early retirement. The trial court stated: "A plain reading of the present statute [Title VII] teaches us that retirement plans which treat men and women differently with respect to their ages of retirement are prohibited." 5 The District Court attempted to decisively resolve the sex discrimination issue: "The Connecticut retirement plan constitutes an unlawful employment practice, because it discriminates between men and women with regard to employment fringe benefits and is therefore found by the Court to be illegal and in violation of federal statutory law." I
Other appellate court cases appear to strongly support this interpretation of Title VII as it applies to retirement systems generally. Bartmess v. Drewrys U.S.A., Inc.,' and Peters v. Missouri Pacific Rail road Co.8' find violative of section 703 (a) a collectively-bargained retirement plan that mandates retirement for women at age 62 and for men at age 65. Rosen. v. Public Service Electric and Gas Co." holds that a retirement plan which pays women who elect early retirement a higher benefit than similarly situated men to elect early retirement, and which mandates that women retire at age 65 and men at age 70, violates Title VII. In Chastang v. Flynn and Emrich Qo.,10 the Fourth Circuit affirms the district court's holding that a vesting schedule which results in female early retirees' retaining a higher percentage of vesting than similarly situated male early retirees represents prohibited discrimination on the basis of sex.
The issues involved in these appellate court cases, together with an additional important issue, are present in Manhart v. City of Los Angeles, Department of Water."' Manhart involves a contributory define benefit pension plan maintained by the Department of Water and Power, City of Los Angeles. Under the Department's plan, men. and women similarly situated received equal retirement benefits. But using
1 42 U.S.C. 2000e--2(a) (1970).
2 Equal Employment Opportunity Act of 1972, Pub. L. No. 92-261, 86 Stat. 103 (1972). 3Fitzpatrick v. Bitzer, 427 U.S. 445 (1976).
*390 F. Supp. 278, 287 (D. Conn. 1974); see also, Peters v. Missouri Pac. R. R. Co., 483 F. 2d 490 (5th 1Mr.
1973), cert. denied, 414 U.S. 1002 (1973).
* Id. at 288.
7 444 F. 2d 1186 (7th Cir. 1971), cert. denied, 404 U.S. 939 (1971). 9 483 F. 2d 490 (5th Cir. 1973), cert. denied, 414 U.S. 1002 (1073).
9 477 F. 2d 90 (3d Cir. 1973).
10541 F. 2d 1040 (4th Cir. 1976).
It 553 F. 2d 581 (9th Cir. 1976), rehearing and rehearing, en bane denied, 553 F. 2d 592 (1977); cert. granted, 46 U.S.L.W. 3214 (1977) (No. 1810).






24

the rationale that women as a group live longer than men, the plan required a 15 percent higher contribution by the female participants. Hence a women participant contributed more than a male participant, yet both received the same benefit.
In striking down this aspect of the pension plan, the Court of Appeals states:
To require every individual woman to contribute 15 percent more into the retirement fund than her male counterpart must contribute because women "on the average" live longer than men is just the kind of abstract generalization, applied to individual women because of their being women, which Title VII was designed to abolish. Not all women live longer than all men, yet each individual women is required to contribute more, not because she as an individual will live longer, but because the members of her sexual group, on the average, live longer.12 The issue in Manhart regarding whether equal benefits as well as equal contributions are required is reflected in the relevant administrative actions occurring in this period. An E.E.O.C. guideline of 1972 states: "It shall be an unlawful employment practice for an employer to have a pension or retirement plan which establishes different optional or compulsory retirement ages based on sex, or which differentiates in benefits on the basis of sex." 13 The same E.E.O.C. regulation deems it violative of the law for a plan to pay unequal benefits to males and females even if employer contributions are equal. On the other hand, regulations issued by the Department of Labor 14 under the Equal Pay Act are to the contrary, and permit unequal benefits so long as employer contributions are equal.
Were the case law to end at this point, the meaning of Title VII in the context of retirement plans generally, and governmental retirement systems specifically, could be viewed as taking shape in a consistent, albeit slow, and, in terms of the specific kinds of discrimination raised, somewhat unpredictable manner.
The Supreme Court's recent decision in General Electric Co. v. Gilbert,15 however, significantly clouds this entire interpretive process.
The facts and holding of Gilbert are well-known and are accordingly not discussed here. The significance of Gilbert is impossible* to predict, both in terms of Title VII law generally and sex discrimination in employer-sponsored retirement plans specifically. Clearly, some employment practices that formerly were believed to be prohibited because discrimination on the basis of sex was involved are now permitted. For example, the Supreme Court in Nashville Gas Co. v. Satty 16 held an employer's policy of forfeiting an employee's seniority following her pregnancy leave violates Title VII, but that the employer's practice of not awarding sick-leave pay to pregnant employees is not a per se violation of the Act. Whether any of the discriminatory pension plan practices struck down in Rosen v. Public Service Electric and Gas Co.,7 (hastang v. Flynn and Emrich Co.,18 Bartrness v. Drewrys U.S.A., Inc.,9 or Peters v. Missouri Pacific Railroad Co.,20 are permitted remains to be seen. The 9th Circuit Court of
12 Id. at 585.
13 29 C.F.R. I IC04.9 (1976).
14 Z) C.F.R. 800.116 (1976).
1 429 U.S. 125 (1976).
16 46 U.S.L.W. 420 (1977).
17 Supra, no0te! 9.
1s Supra, note 10.
444 F. 2d 1186 (7th Cir. 1971), cert. denied, 404 U.S. 939 (1971). 20 Supra, note 8.





25

Appeals, in affirming its pre-Gilbert decision in Alan art v. City of Los Angeles Department of Water,2' concludes that the illegality of the practice of the Los Angeles Department of Water and Power in requiring a 15 percent higher contribution from female participants is not made permissible by the Supreme Court's decision in Gilbert. Importantly, however one of the three judges who issued the original 9th Circuit opinion and considered the petition for rehearing states very strongly in a dissent to the denial of the motion for a rehearing that appellants (City of Los Angeles), because of Gilbert, should have at least had the opportunity to prove that its retirement plan was justified. Further, Judge Kilkenny continues in dissent, the sex discrimination aspects of the Los Angeles plan are exactly analogous to those aspects of General Electric's disability plan, and hence should be permitted in light of Gilbert.*2
While the 9th Circuit Court concluded that the Supreme Court's decision in Gilbert does not make legitimate the Los Angeles Water Department plan, Judge Kilkenny's dissent signifies that Gilbert will alter the analysis of retirement plans that has developed over the years in the context of Title VII.23 The precise way in which this developing regulatory and case law will change remains to be seen.
The Supreme Court has granted certiorari in Manhart,24 and its treatment of the appeal may resolve many of the issues sharply disputed by Judge Kilkenny as well as the broader issues relating to the use of gender based mortality tables.
Legislative efforts presently underway to reverse Gilbert will, if successful, clarify that the kind of coverage in the General Electric .disability plan represents the kind of discrimination on the basis of sex prohibited by Title VII.25 It is not expected that these pending legislative proposals will address the variety of practices described in Bartmess v. Drewrys U.S.A., Inc., Peters v. Missouri Pacific R.R. Co., Rosen v. Public Serv.ie Electric and Gas Co., and Chastang v. flynn and Emrich Co.26
Government plans, in any event, will clearly remain subject to Title VII, and thereby subject to a pervasive and significant from of federal regulation.
AGE DISCRIMINATION IN EMPLOYMENT ACT
The Age Discrimination in Employment Act I respresents an additional aspect of federal law that has a significant impact on state and local public employee retirement systems.
The Act, as passed in 1967, prohibits employers from discriminating against individuals on the basis of age with respect to hire, discharge, or other terms and conditions of employment, 2 if those individuals are between the ages of 40 and 65.3 Exceptions from the general proscription against discrimination on the basis of age include the well known bona fide occupational qualification situation,' bona fide
21 Supra, note 11.
2- 1,J. at 594.
23 See, e.g., Mitchell v. Bd. of Trustees of Pickens County School Dist., Civ. No. 75-143 (D.S.C. July 27,
1977).
24 Supra, note 11.
25 S: 995, 95th Cong. 1st Sess. (1977); H.R. 6075, 95th Cong. 1st. Sess. (1977). 26 Upra, notes 17-20.
129 U.S.C. 621-634 (1979).
2 Id. 623.
3 Id. 631.
4 Id. 623(f) (1).






26

seniority system or employee benefit plan exceptions,5 and the good cause exception.6
The 1967 Act specifically excluded all federal, state, and local government entities from the definition of employer.7 For state and local governments, the key Act is the 1974 Act amending the A.D.E.A.81 State and local governments were specifically brou ght within the definition of "employer",9 and hence are treated equally with private employers under the Act.'0 The prohibited employer practices, as well as the exceptions permitted in section 623 (f), were not altered by the 1974 amendments.
Notice must also be taken of pending amendments to the Age Discrimination in Employment Act, H. R. 5383, as pass 'ed by the House of Representatives on September 23, 1977, would (1) raise the upper limit of the protected age group for private sector and state and local government employees from age 65 to 70, and, most important for present purposes, (2) clarify that involuntary retirement before age 70 on account of age shall not be required or permitted by any seniority system or employee benefit plan." The Senate passed H.IR. 5383 on October 19, 1977, but exempted from the prohibition against involuntary retirement at age 65 executives entitled to large pension benefits and fully tenured college and university professors. As is immediately apparent, the enactment of H.R. 5383 would significantly alter the A.D.E.A. as it relates to public employee retirement systems.
The Age Discrimination in Employment Act, it is recalled, does permit an employer, including a governmental employer, to require an employee to retire, or otherwise discriminate against an employee because of age, if the employer does so
To observe the terms of a bona fide seniority system or any bona fide employee benefit plan such as a retirement, pension, or insurance plan, which is not a subterfuge to evade the purposes of this chapter, except that no such employee benefit plan shall excuse the failure to hire any individual . *1
A quick reading of this provision demonstrates its inherently contradictory nature: an employer may not discriminate on the basis of age unless the employer maintains a bona fide pension plan which permits or requires discrimination on the basis of age. As might be expected, the meaning of 29 U.S.C.A. 623 (f) (2), better known as the "4 (f) (2) exception",'" has not been clear to the courts either.
The varying analyses that have surfaced with regard to the 4 (f) (2) exception have been substantially clarified in recent months. In Mc~fann v. United Air Lines,'" the Supreme Court holds that a
Id. 632 (f) (2).
SId. 623(M(3).
7 Id. 630(b).
8 A.D.E.A. Amendments of 1974, Pub. L. No. 93-259, 88 Stat. 74.
9 29 V.s.C. 630(b) (Supp. V 1975).
10 This section does not include a discussion of the application of the A.D.E.A. to the Federal Government as an employer cf. Christie v. Marston, 551 F. 2d 1080 (7th Cir. 1977).
It H.R. Rep. No. 527 Part 1, 95th Cong., 1st Sess. (1977).
13 Supra, note 5 (emphasis added).
129 U.S.C. 623(f) (2) (1970) (corresponds to Age Discrimination in Employment Act of 1967, section 4ff)
(2)).
1446 TJS.L.W. 4043 (1977).





27

retirement plan which antedates the enactment of the A.D.E.A. in 1967 and requires retirement prior to age 65 is not prohibited by the Act, but rather is permitted under the 4(f)(2) exception. The Court extensively reviews the legislative history of section 4(f) (2), and concludes that it was "intended to permit observance of the mandatory retirement terms of bona fide retirement plans, but that the existence of such plans could not be used as an excuse not to hire any person because of age." It also holds that "we find nothing to midicate Congress intended wholesale invalidation of retirement plans instituted in good faith before its passage, or intended to require employers to bear the burden of showing a business or economic purpose to justify bonafide pre-existing plans.... This decision by the High Court indicates that the analysis of the 5th Circuit Court of Appeals in Brennan v. Taft Broadcasting Co.,16 in which a pre-1967 retirement plan requiring retirement at age 60 is found permissible, is correct, thereby resolving the clear conflict that had previously existed between the 4th and 5th Circuits with regard to the legality of mandatory retirement policies established before 1967.
An alternative analysis that the Supreme Court has decided not to consider is found in the 3d Circuit's opinion in Zinger v. Blanchette.7 The Court focuses not on whether the establishment of the plan and its pre-age 65 mandatory retirement policy precedes or follows the enactment of the A.D.E.A. in 1967, but rather on whether the retirement plan, and particularly its benefit level, is bona fide. The Supreme Court's denial of the petition for certiorari in Zinger, coupled with the Court's decision in Mc2lann, indicates that pension plans with mandatory retirement features prior to age 65 fall within the 4(f) (2) exception if the plan was either (1) established prior to the enactment of the A.D.E.A. in 1967, or (2) established subsequent to 1967 but contains a benefit schedule sufficiently large to make the plan bona Jide. Whether a post-1967 plan mandating retirement prior to age 65 but containing a very meager benefit schedule is permitted under the 4(f) (2) exception remains to be seen.
Additional issues are present when the employer involved in an A.D.E.A. action is a state or local government acting as an employer. This Report, supra, discusses at length the constitutional limitations of Congressional power in the context of the federal relationship. With regard to the applicability of the A.D.E.A. to state and local governments in light of National League of Cities, et al., v. Usery,'8S: it appears that federal law prohibiting employment discrimination by state and local governments, and having only a slight impact on the fiscal operations of the regulated government entity, is not prohibited by National League of Cities. A number of courts have so held in recent months.19 The precise significance of National League of Cities,
15 Id. at 4046.
16500 F. 2d 212 (5th Cir. 1974).
17 549 F. 2d 901 (3d Cir. 1977), cert. denied, 96 U.S.L.W. 3436 (1978).
Is 426 U.S. 833 (1976).
19 See, e.g., Usery v. Bd. of Educ. of Salt Lake City. 421 F. Supp. 718 (D. Utah 1976); Usery v. Bettcndorf Community School Dist., 423 F. Supp. 637 (S.D. Iowa 196); Christensen v. Iowa, 417 F. Supp. 423 (N.D. Iowa 1976); Usery v. Allegheny County Inst. Dist., 544 F. 2d 148 (3d Cir. 1976). cert. denied, 45 U.S.L.W. 3651
(1977); Usery v. Charlestown County School Sys., No. 76-2340 (4th Cir. July 25, 1779); Usery v. F ort, AMisoa : Community School Dist., No. C 75-62-1 (S.D. Iowa, motion to dismiss denied Sept. 1, 1976).; Usory v. Sioux City Community School Dist., No. C 76-4024 (N.D. Iowa. motion to dismiss denied Aug. '-0, 1976); Riley v. Univ. of Lowell, Civ. No. 76-1118-M (D. Mass., motion to dismiss denied July 22, 1976).


74-365-78 3





28

and alternative bases of Congressional power, are discussed above.2" A related issue involves the ability of federal courts to award money damages to be paid by a state government, a clear possibility under the Act.2I As is discussed at length earlier in this report,2 the 11th Amendment issue just described is resolved in favor of federal court jurisdiction to award money damages against a state once it is appreciated that the Age Discrimination in Employment Act is enacted pursuant not only to Congress' power under the Commerce Clause,2 but also an exercise of Congressional power under section 5 of the Fourteenth Amendment.
SECURITIEs ACTS
Existing federal law regulating the securities markets represents an ongoing federal involvement in state and municipal pension funds.
It is well established that the antifraud provisions of the Securities Act of 1933' and the Securities Exchange Act of 1934 2 apply to issuers of state and local government securities.3 That is, state and local government issuers must disclose to potential investors all information that might reasonably be useful to potential investors in reaching an investment decision. As has often been stated, the focus is on materiality. That which is material to the investment decision must be fully and fairly disclosed to the prospective purchaser, if liability under the antifraud provisions of the '33 and '34 Acts is to be avoided. "A fact is material if it concerns information about which an average prudent investor ought reasonably to be informed before purchasing the security."
The Securities Act Amendments of 1975 5 did not alter this basic regulatory scheme. These amendments do significantly increase federal regulation of the municipal securities markets. But regulation of the issuer directly is not changed. State and municipal issuers remain subject to the antifraud provisions of the Acts, and remain exempt from predisclosure registration and other Securities and Exchange Commission pre-sale regulation.
Despite lengthy experience with antifraud actions based on the '33 and '34 Acts, the precise role that information relating to an issuer's pension plan plays in full and fair disclosure of all material information is far from clear. Few actions have been brought under the Securities Acts in which the plaintiff has claimed a governmental issuer failed to meet the antifraud disclosure requirements, thereby causing injury to the prospective purchaser.6 There are no significant reported decisions in which such non-disclosure or misleading disclosure involved the municipal or state issuer's retirement systems.
In conclusion, it appears that federal law regulating the issuance and sale of securities, in theory, is applicable to state and municipal issuers of securities. The precise significance of this form of federal regulation, specifically in terms of the kind of information relating to a pension plan which a governmental issuer must disclose, has not been clarified either by statutory or case law.
2o Supra, p. 17.
-' 2) U.S.C. 26 (1970).
Supra, p. 16.
U.S. CONST. art T. 8, ci.3.
115 U.S.C. 77(Q) (1971).
15 U.S.C. 78 j(1) (197G).
SEC v. Chrles A. Morris and Assoc. Tic., 386 F. Supp. 1327 (W.D. Tenn. 1973).
4 Johis Ilopkins Uiiiv v. Itutton, 422 F. 2d 1124 (4th Cir. 1970, cert. denied., 416 U.S. 916 (1974).
6 Pub. L. No. 94-2.9, K Sta 97 (1975).
6,*e, e.q., Yeonans v. Kenitucky, 514 F. 2d 993 (6th Cir. 1975); SEC v. Washington County Utility Dist., No. 2-77-15 (E.D. Tenn. 1977).






29

A second and extremely topical securities law involvement in pension and Nvelfatre benefit plans, theoretically including public employee retirement systms, centers on the recent decision by the Seventh Circuit Court of Appeals in Daniel v. International Brotherhood of Team.gterS.7 In Dniel, the Court of Appeals affirmtd the holding of the District Court that mandatory participation by an employee in a noncont.ibutory pension. plan represents the sale of a security by the plaf to the, participant, thereby making applicable to the plan the antifraud (disclosure) provisions of the Securities Acts."8 Presumably a Daniel theory is more feasible in the context of a contributory plan, in that it is more easily appreciated that the decision to enter into an employment relationship represents an investment decision by the prospective employee in terms of the. pension plan.
The significance of the Seventh Circuit's decision is still far from clear. Review has been granted by the Supreme Court,' and a number of amici curiae urging reversal seem likely.' Additionally, a number of district courts have reached decisions contrary to Daniel on similar facts." Furthermore, Daniel itself is still at a preliminary, albeit vital, stage. The present decision by the Seventh Circuit and its appeal to the Supreme Court involve denials of defendant's motions to dismiss the plaintiff's complaints on the ground that the trial
court lacked subject matter jurisdiction and that plaintiffs failed to state a claim upon which relief could be granted; that is, that the Securities Acts create no cause of action based upon the non-disclosure which plaintiff has alleged. Thus no finding of liability has been made, and the courts have not yet had the opportunity to consider what represents "materiality" in the context of the purchase of a security in the form of participation in a plan, when any such disclosure must be made, what standard (e.g. scienter, negligence) will be used to, determine liability, and so on.
It should alsQ. be hoted that the Securities and Exchange Cormmission has recently taken an increasingly expansive view of securities law and fraud in the context of pension and welfare benefit plans. In S.E.C. v. Shenker, 1 the U.S. District Court for the District of Columbia permanently enjoined defendants from continued violations of the antifraud provisions of the Securities Acts. Two of the counts in the S.E.C. complaint, upon which the permanent injunction and a related consent decree were based, charge that certain trustees of a pension fund and a welfare fund violated the antifraud provisions of the Securities Act by failing to disclose their participation in a course of conduct through which assets of the employee benefit plan trust funds were used for the benefit of persons other than the participants: in the funds. It is apparently the view of the S.E.C. and the District Court iss"ing th~ ifrjunct ions and approving the consent decrees in
561 F. 2d 1223 (7th Cir. 1977) Cert. granted, No.77-753, Feb. 21, 1978.
8 *Stipts, notes 1 shd 2. .
No. 77i-753 Feb. 21.4978
SoAmici curiae before the 7th Circuit included: SEC, Gray Panthers, Institute for.Public Interest Hceresentation. Teamsters for a Democratic Union, Secretary of Labor, E RISA Industry Committee, and the National Coordinating Committee for Muliiemployer Plans.
1 Hurnv. Retirement Trust Funds, Etc., 424 F.Supp.80 (C.D. Cal. 1976); Weins v.International Brotherhood of Teamsters, C.A. Number 76-2517H, (C.D. Cal. 1977): Robinson v. United Mine Workers of Ameha Health and Retirement Funds. Number 77-0698 (D.D.C. 1977).
1 No. 77-1756 (D.D.C. 1977).






30

S.E. C. v. Shenker that fiduciary breaches by pension and welfare plan officials, if not disclosed, represent violations of the Securities Acts. 13
Unresolved issues involving astronomical sums abound for governmental pension and welfare plans if Daniel is upheld. The entire concepts of (1) materiality ill the context of a governmental plan issuer and (2) a plan participant or prospective participant as a purchaser of a security in the form of an interest in the pension or welfare plan, axe too novel to be considered with any confidence or certainty. Standards which will determine retroactive liability, collective bargaining considerations, and numerous other- issues will have to be addressed. The tremendous concern expressed by plan sponsors, and others in the private sector, following Daniel, is equally relevant in the context of governmental plans. If Daniel is upheld, or not firmly reversed, and the Securities and Exchange C mission continues in its increasingly aggressive effort to apply the antifraud provisions of the Securities, Acts to pension and welfare plans, public employee retirement systems will be faced with a federal statutory, regulatory, and enforcement framework of considerable significance.

INTERNAL REVENUE CODE
The Internal Revenue Code of 1954 1 is of obvious significance to public employee retirement systems. The degree to which pi blic employee retirement systems are subject to various Code provisions has a direct impact on virtually every aspect of a state or local government employee benefit plan.
Before considering to what extent public employee plans are subject to the qualification requirements of the Internal Revenue Code,' the result of receiving "qualified" status, and the relevancy to state and local government plans, should be noted.
Briefly stated, qualification of a pension plan under section 401 (a) of the Internal Revenue Code results in three major tax benefits for employees, employers, and their pension plans:
(1) The employer's contributions to the plan are deductible
when made, even if the employee is not vested in them at that
time; 1
(2) the earnings of the pension trust funds are not taxed
currently ; 4
(3) the contributions made by an employer to a plan on behalf
of an employee are not currently imputed to the employee for income tax purposes, even if vested.5 Also, advantageous tax treatment is afforded to a participant who receives a lump-sum
6
distribution from a qualified plan, and favorable income tax treatment 7 and estate tax treatments are available with regard
to death benefits paid from a qualified plan.
is Compare Santa Fe Industries, Inc. v. Green, 430 U.S. 462, (1977), in which the Supreme Court holds that fiduciary breaches in the absence of nondisclosure are not actionable under the antifraud provisions of the secarit ies, aets.
I. R.C. I et soq.
1. R.C. 401 et seq.
id. 1404.
Id. 501 (a).
Id. 402(a).
Id. 402(e).
Id. 101 (b).
Id. 1 2039(c).





31

A number of observations are in order with regard to state and local government plans and the tax implications of characterizing a retirement plan as qualified. Because gross income does not include income accruing to a state or local government,9 the advantage to the employer in having a qualified plan-immediate deductibility of employer contributions, even if not vested in the employees-is not relevant. Whether the second tax advantage to qualified statusdeferral of taxation of the income of the pension trust-is of relevance to governmental plans is unclear. Where the trust is part of the governmental entity,10 trust income presumably would be imputed to the government and tax exempt on that basis. With minor exception, the issue has not been litigated or addressed administratively by the Internal Revenue Service.
It is clear that the third advantage of qualified status-deferral of recognition as current income by the employee of employer contributions in which the employee is vested-is relevant for governmental plans. Deferral of recognition of income is a matter of grace on the part of the federal government,11 and the fact that the income in this instance is derived from a governmental employer is of no consequence. If the governmental plan is not treated as qualified, the plan participant must currently recognize as income the contributions made by the employer, to the extent the participant is vested in the contributions.2 If the governmental plan is treated as qualified, then the participant may defer recognition of the value of his vested, funded interest in the plan until it is actually distributed to him.3 Such actual distribution of the participant's plan interest And recognition of income will probably occur after the participant is no longer an active employee. A considerable tax savings may result in that the participant may recognize the income in a year in which taxable income is considerably less than the year(s) in which the participant became vested in the employer's contribution. Also, as in the private sector, the plan participant of a qualified governmental plan is eligible to receive favorable tax treatment if his interest is distributed in the form of a lump sum distribution.4
It is clear that almost all of the substantive qualification requirements added to the Internal Revenue Code by ERISA specifically exempt governmental plans from coverage., A governmental plan
is defined as "a plan established or maintained for its employees by the . government of any State or political subdivision thereof, or by any agency or instrumentality of any of the foregoing.. Such governmental plans are specifically exempted from the Code provisions added by ERISA which relate directly to participation, '
171
vesting, funding,'8 prohibited transactions,9 joint and survivor annuities,20 plan merger and consolidation,2' alienation and assign'Id. 115.
10 See, e.g., Fitzpatrick v. Bitzer, 427 U.S. 445 (1976).
!,New Colonial Ice Co. v. Helvering, 292 U.S. 435 (1934). 12 I.R.C. 402(b).
1s Id. 402(a).
14 I1. 402(e).
is Id. 414(d).
16 Id. 410(c) (1) (A).
17 Id. 411(e) (1) (A).
1I Id. 412(h) (3).
I Id. 4975(g) (2).
20 d. 401(a) (1t).
21 Id. 401(a) (12).





32

-rofnt of plan benefits, payment of benefits,2 certain Social Security benefit increases,2- and withdrawal of employee contributions." However, governmental plans were not exempted from the ERISA Internal Revenue Code provisions placing specific limitations on benefits and contributions under qualified plans.8
Importantly, qualified governmental plans appear to remain subject to the Internal Revenue Code participation, vesting, funding, and fiduciary standards applicable to all qualified plans before the enactment of ERISA. Code section 410, while exempting governmental plans from the new participation standards,7 states that such plans do remain subject to the pre-ERISA participation standards.8 Thus to be qualified, governmental plans must in theory cover 70 percent of all the employer's employees, or 80 percent of all employees who are eligible to benefit in the plan if 70 percent or more of all the employees are eligible to benefit, and employees who are employed more than 20 hours per week for at least five months per year. In addition, such plans apparently must not have service requirements in excess of five years, or otherwise discriminate in favor of highly compensated employees.29 Similarly, in the vesting area, ERISA adds Code section 411, which exempts governmental plans,0 but preserves the proERISA Code vesting requirements for governmental plansY4 Thus governmental plans to be qualified must provide for vesting (1) upon a participant's attaining normal retirement age or the age and service requirements contained in the plan, (2) upon plan termination, and ,(3) to generally prevent discrimination in favor of highly compensated
-employees, 2
With regard to funding, Code section 412, added by ERISA, exempts governmental plans from its scope,43 and preserves the preERISA safe-haven rule relating to whether a plan termination has occurred upon suspension of contributions by the plan sponsor." Generally, ERISA did not change the pre-ERISA "funding" require, meats applicable to governmental plans. Prior to the enactment of ERISA, however, in the event of the complete suspension of contributions under a governmental plan, an IRS "safe-haven" rule applied wherein the plan would not be considered "terminated" if contributions equal to at least normal cost plus interest on the unfunded liability had been made (thus preventing the initial unfunded liability from increasing) .5 This safe-haven rule continues to apply to governmental plans in the post-ERISA period. In the event a governmental pan, subject to the Internal Revenue Code, terminates or is deemed to e terminated by the Internal Revenue Service, plan assets must be allocated to those benefits in which plan participants are vested (a provision primarily intended to prevent a reversion of plan assets to the plan sponsor) ,36
?2 Id. 401 (a) (13).
93 Id. 401(a)(14).
I Id. 401 (a) (15).
25 Id. 401(a)(19).
26 Id. 415.
27 Supra, note 16.
28 I.R.C. 410(c)(2).
20 Id. 401(a)(3) and (4) (Sept. 1, 1974).
20Id. 411(e)(1)(A).
31 Id. 411 (e) (2). 22 Id. 401 (a) (4) and 401(a)(7) (Sept. 1, 1974).
33 Supra. note 18.
4 11.R. Rep. No. 93-1280, 93d Cong., 2d Sess. 291 (1974). 35 Treas. Reg. 1.401-6(c)(2) (Sept. 1, 1974).
30 J.R.C. 401(a)(7) (Sept. 1, 1974); I.R.C. I 412(h).





33

With regard ta fiduciary standards, ERISA adds Code section 4975, but exempts governmental plans from its scope." Pre-ERISA standards are maintained for governmental plans, however.38 Thus, in theory, a qualified governmental plan may not lend any part of its income or corpus to a creator of the trust, substantial contributor to the trust, or similar entity, without adequate security. The plan also may not pay excessive compensation to such a person, purchase securities at excessive prices from such an entity, sell securities at an undervalued price to such an entity, etc.39
Until recently (see next paragraph), Internal Revenue Service Rulings suggested that governmental plans are subject to these various qualification requirements of the Internal Revenue Code.40
This interpretation by the Service, it should be carefully noted, has had virtually no practical significance to state and local plans. Enforcement of the qualification standards against public plans has been for the most part non-existent. No plans have been disqualified, and it does not appear that the Service has ever successfully imputed to a public plan participant the value of his vested, funded, pension plan interest in a situation in which the participant sought to defer recognition of that income; i.e. receive the benefit of having the plan viewed as quialified.
Furthermore, following a recent flurry of protest when the Service did begin to apply the qualification requirements to state and local governmental plans, the Service announced 41 that it will reconsider whether (1) the qualification requirements apply to public plans, and
(2) the trusts of such plans are subject to tax on their income. Until such a review is completed, the Internal Revenue Service has indicated it will resolve these issues in favor of the taxpayer or governmental unit; that is, continue to treat the plan as if it were qaalified.0 Any attempt to explain this attitude on the part of the Internal Revenue Service is, of course, highly speculative. It does appear, however, that the exstene of a qualification letter for the retirement plan for federal judges 4--a plan thought t9 be highly discriminatory in favor of highly compensated empoyees- has played some role in the decision not to challenge the status of governmental plans.
Of interest and importance in the context of governmental plans
-ad the Internal Revenue Code is P.L. 94-236, enacted by Congress in 1976 in response to the New York City fiscal crisis. A key elemen, of the refinanemig program for New York City in 1976 involved the purchase by five New York City pension funds of New York City bonds. Great concern was expressed by the trustees of the pension plans, the Internal Revenue Service, and the Congress that the proposed investments violated the exclusive benefit rule of Code section 401(a) and certain fiduciary provisions of Code section 503(b), as those provisions apphe4 to governmental plans (see above). P.L.
V Supra, note 19.
39Id. 503(b).
49 Rev. T$ul. 72-1%7i92-1 C.] 10w
41 T. R. 1869, Aug. 10, 1977.
42 Id.
43Rev. Rul. 61-218, 1961-2 C.B. 102.
4490 Stat. 238 (94th Cong. 2d Sess., 1976).





34

94-236 waived the prohibited transactions and exclusive benefit rules of the Internal Revenue Code with regard to the. purchase of New York City securities by the five city pension funds, a step necessitated by the general unmarketability and high risk nature, of such bonds at that time. The Ways and Means Committee, in approving the bill, stressedl that it should not be interpreted as a precedent under which governments can use plan assets to assist cities in raising revenuesduring periods of financial crisis, even though such use of plan assets. might violate the exclusive benefit or prohibited transaction rules of the Internal Revenue Code .45 Despite the fact that those Code provisions which are applicable to governmental plans are generally not enforced by the Internal Revenue Service, the presence of such provisions of law serves as a significant form of federal regulation of governmental pension plans, as the history of P.L. 94-236 demonstrates.
Much recent activity ha-s focused on the issue of whether state and local government plans are subject to various reporting requirements contained in the Internal Revenue Code. The annual registratration statement is limited to plans subject to the vesting requirements of ERJSA.6 Inasmuch as governmental plans are exempted from those requirements,7 they are similarly exempted from the annual registration statement. The annual return requirement contained in the Code'48 however, does not include an exemption for governmental, plans, and the Internal Revenue Service has taken the view that governmental plans are subject to the annual report requirement .41 Penalties for failure to file the annual report are $10 per day for the period in which the report has not been filed, and the total penalty may not exceed $5,000.50,
In summary, pre-IERISA Internal Revenue Code provisions relating, to retirement systems generally apply to state, and local government plans. Certain Code provisions added by ERISA, such as those relating to limitations on benefits and contributions, and annual returns, also, apply to governmental plans, although it is clear that the Internal Revenue Code provisions -added by ERISA relating to participation, vesting, funding, prohibited transactions, and so on, are not appli-, cable. Tfhe significance of Internal Revenue Code regulation of state and local governmental plans has not been appreciated largely because the Internal Revenue Service has been lax. to enforce the Code in the context of state and local government retirement systems. The recent formalization of this non-enforcement policy jin I.R. 1869 "' indicates that it will continue for at least the immediate future.
In those few instances in which the Service, has enforced or threatened to enforce qualification requirements against public plans,, the impact on the plans involved has been quite significant. A pensionplan for the police and firemen in St. Joseph, Missouri, for instance,, paidI several thoulsand1 dollars in "income taxes"~ on earnings of the pension trust to the Internal Revenue Service in the early 1970's. It is clear thalt a vioyorous effort by the Internal Revenue Service to enforce the% penalties upon failure to file the section 6058 annual
45 H.R. Ron. No. 94-851, 94th Cong., 2d Sess. (1976).
01. H.C. 6057 (a) (1).
47 Supra, riot (3 17.
48 1. R.C 6058.
49 T. I. 1798, Apr. 21, 1977.
LI .C 6058(d) and 6652(f).
61 Sapra, note 41.





S35

reprtform would have a. major impact on public employee retirement systems. As with so many other federal laws already enacted, the Internal Revenue Code represents an extensive and significant form of federal reg-ulation of public employee retirement systems, the significance of which has not been realized largely because the govern-ment has chosen not to enforce the relevant statutory provisions.

SOCIAL SECURITY
The federal Social Security Program'I is of obvious significance to
-state and local government retirement systems. Generally, under existing law, employees of state and local governments are covered by Social Security* only if the government unit and the Secretary of Health, Education, and Welfare have entered into voluntary agreements to provide such coverage.2 There are a number of important variations on this voluntary theme .3 For instance, a voluntary coverage agreement may not be placed in effect for employees presently included in a state or local retirement plan unless a majority of eligible employees, in a secret, written, ballot referendum, approve Social Security coverage.4
A state or local government may terminate coverage for a group of employees by giving notice two years in advance, after the coverage has been in effect for five years. The notice of desire to terminate may be withdrawn before the expiration of the two year notice period. Once coverage has been terminated for a group of employees, it can never be reinstated for that group."
At present, approximately 70 percent of all state and local governmient employees are subject to 'Social Security coverage.6 Recalling that the governmental employer and the covered employee each contributes 6.05 percent of the employee's gross earnings to Social Security, it is clear that whether a group of public employees is subject to
-Social Security "coverage will dramatically affect the availability of funds to be contributed to the state or local government retirement system. Benefit levels, integration formulae, variety of benefits, and so
-on of the state or local government pension plan will obviously be vitally affected by whether the employees covered by the plan are also covered by Social Security.
The termination of coverage provisions of Social Security is o f 'increasing significance to public employee retirement systems. A large number of public employee groups, representing over 350,000 workers, submitted notices of intent to terminate Social Security coverage in the period from 1973 to 1975.1 A number of these plans withdrew their notices of intent to terminate coverage prior to the expiration of the two year notice period, presumably concluding that comparable benefits could not be obtained elsewhere at a comparable cost.8
1 42 U.S.C. 301 et seq. (1970).
2 4-9 U.S.C. J 418 (1970).
SId.
4 Id.
6 See Part TV, Chapter B.
Statement of James B. Cardwell, Commissioner of Social Security, reported in 151 ENA Pension
-Reporter A-1.






36

The enactment of the Social Security Financing Amendments of 1977 9 changes the Social Security system in a number of extremely significant ways. Contributions to the Social Security system by both employers and employees will rise dramatically, in terms of both the rate itself and the maximum earnings subject to the tax. As the costbenefit formula changes, it can be expected that a number of governmental plans will re-assess their benefit structures and participation or non-participation in Social Security.
Extremely significant changes in the Social Security system with direct implications for federal, state and local retirement systems are presently under consideration by the federal government. In accordance with Public Law Number 95-216, the Secretary of Health, Education and Welfare is required to undertake a study and report on mandatory coverage under Social Security of employees of federal, state and local governments in consultation with the Office of Management and Budget, the Civil Service Commission, and the Department of the Treasury. The study is to examine the feasibility and desirability of coverage of these employees and is to include alternative methods of coverage, alternatives to coverage, and an analysis under each alternative of the structural changes which would be required in retirement systems and the impact on retirement systems benefits and contributions for affected individuals. The report to be made to the President and the Congress is due two years after enactment.

REVENUE SHARING
It is clear that a great many state and local governments use portions of monies received under the State and Local Fiscal Assistance Act I to fund governmental pension plans2
As passed in 1972, the Federal Revenue Sharing Act contained few limitations on the ways in which states in receipt of revenue sharing funds could use such funds. The 1972 Act, however, did require that units of local government receiving such funds use them only for "priority expenditures", defined to mean "only(1) ordinary and necessary maintenance and operating expenses
for
(A) public safety (including law enforcement, fire protection, and building code enforcement),
(B) environmental protection (including sewage disposal,
sanitation, and pollution abatement),
(C) public transportation (including transit systems and
streets and roads),
(D) health,
(E) recreation,
(F) libraries,
(G) social services for the poor and aged, and
(H) financial administration, and
(2) ordinary and necessary capital expenditures authorized by
law." I
Pub. L. No. 95-216 91 Stat. 150 (1977).
1 31 U.S.C. 1221-12h4 (Sup. V 1975).
2 The Pulic Stir F'mploeP' Rdtirimnt Income Security Act of 1976: Hearings on H.R. 9165 e d 1.R. 80 before the Subcomn. on Labor Standards of the House Comm. on Education and Labor, 94th Cong., 1st Bess. (1975) (Statement of John M. Miliron).
3 U.S.C. 1222 (Supp. V 1975), (repealed by State and Local Fiscal Assistance Amendments of 1976, Pub. L. No. 94-488 3(a), 90 Stat. 2341).






37

The absence of any further refinement of what constitute "priorityexpenditures" led some observers to conclude that funds could be used by local governments for virtually any purpose by reallocating locallyraised revenues from "priority categories" to "non-priority categories",
)) 4
and using the federal revenues for "priority categories
The few cases in which proposed local government use of federal
revenue sharing funds was challenged appear to support this view'Given that states under the 1972 Act had unrestricted use of revenue sharing funds, and local governments were limited in permitted uses only to broadly defined "priority categories", use of revenue sharing" funds by state and local governments to fund governmental pension plans appears completely appropriate. It should also be noted that the states and localities are required to report on the use of revenue sharing funds,6 and hence presumably the Congress was fully aware that funds were being used by governments to fund retirement systems.
Finally, whatever doubt might have remained regarding the propriety of local governments' use of federal revenue sharing funds to fund or pay pension benefits was removed by the 1976 amendments to the Federal Revenue Sharing Act.7 The limitation to "priority categories" for local government use of revenue sharing funds was repealed, thus leaving local governments, like state governments, completely free to use federal revenue sharing funds to fund or pay governmental retirement benefits.
FAIR LABOR STANDARDS ACT
The Fair Labor Standards Act affects public employee retirement systems in two ways.
The Act establishes minimum wage and overtime provisions for
employees. The Act as originally passed in 19382 excluded states and political subdivisions thereof from the definition of "employer'". In amendments, to the Act in 1966, the exemption for states and political subdivisions was narrowed, and most employees of state and municipal schools, hospitals, and institutions were brought within the coverage of the Act. The Supreme Court sustained the constitutionality of this extension in -Maryland v. Wirt2,4 holding that "If a State is engaging in economic activities that are validly regulated by the Federal Government when engaged in by private persons, the State too may be forced to conform its activities to federal regulation"2.
The Act was again amended in 19746 to bring within its coverage all state and local government entities. The Supreme Court, in National League of Cities v. Usery 7 (discussed at length, s'upra), declared this extension of coverage to be violative of the Tenth Amendment's reservation of power to the states and hence unconstitutional. The Court also reversed its decision in Maryland v. Wirtz. Thus subsequent to the June, 1976, decision in National League of Cities, employees
4 See, e.g., The Revenue Sharing Act of 1972: Untied and Untraceable Dolars from Washington, 10 Harv. 7. Legis. 276 (1973).
5 Yovetich v. McClintock, 165 Mont. 80, 526 P. 2d 999 (1974); Mathews v. Massell, 356 F. Supp. 291 (N J). Ga. 1973).
6 31 U.S.C. 1241 (Supp. V 1975).
7 State and Local Fiscal Assistance Amendments of 1976, Pub. L. No. 94-488, 00 Stat. 2941. 129 UjS.C. 201 et seq. (1970).
252 Atat. 1060, Chnnter 676 (19MI.
3 Pub. L. No. 89-601, 80 Stat. 830 (1966).
4 R92 U.S. 183 (1968).
&Id. at 197,
Pub. L. No. 93-259, 88 Stat. 55 (1974).
7426 U.S. 833 (1976).





38

of states and political subdivisions generally are not beneficiaries of the protections of the minimum wage and overtime provisions of the Fair Labor Standards Act. The F.L.S.A. remains applicable, however, to those public employees engaged in enterprises not integral to the general functions of the state or local government employer.
To the extent that benefits or funding requirements are based on employees' compensation, state and local governmental plans have been, and to some degree, continue to be, affected by the minimum wage and overtime provisions of the F.L.S.A.
The equal pay provisions of the Fair Labor Standards Act 8 also affect public employee retirement systems. rhese provisions generally rohibit employers covered by the Act from discriminating on the asis of sex with respect to the payment of wages to employees. As discussed above, certain state and local governmental employers were made subject to the Act, including its equal pay provisions, in 1966, and all such employers were brought under its coverage in 1974. A large majority of the courts that have addressed the issue I have held that this aspect of the F.L.S.A. is distinguishable from the Act's minimum wage and overtime provisions in both its impact on basic state and local government functions as well as the jurisdictional basis of its enactment, and is not unconstitutional under the Supreme Court's decision in National League of Cities v. Usery.'0 Thus state and local governments continue to be subject to the equal pay provisions of the F.L.S.A., and state and local governmental retirement systems continue to be affected by the Act to the extent the systems are affected by sex-based variations in the employees' compensation.

LABOR MANAGEMENT RELATIONS ACT
Public employee retirement systems are not affected by the National Labor Relations Act and the Labor Management Relations Act of 1947.1 These Acts establish the right of employees to bargain collectively through representatives of their own choosing, and otherwise engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection. It also establishes a series of prohibitions limiting employer actions which interfere with basic rights guaranteed to employees under the Act. If the L.M.R.A. applies to an employer-employee relationship, the terms of pension and welfare plans, at least for active employees,2 are clearly among the topics which are to be negotiated by the employee's collective bargaining agent and the employer. This feature of the L.M.R.A. has obviously been of extraordinary significance in the private sector.
The L.M.R.A., however, clearly excludes from its definition of employer "any state or political subdivision thereof",3 thereby removing from its coverage the vast majority of what are commonly thought of as public employers. Hence the development of benefit plans by such public employers for public employees is not affected by the Labor Management Relations Act.
S 2) U.S.C. 1 206(d) (1970).
<,, r !., Usry v. Allegheny County Tnst. Dist., 51I F. 2d 148 (3d Cir. 1976), cert. denied, 4.5 U.S.L.W.
3(51 (1977)- Usery v. Dallas Independent School Dist., 421 F. Supp. 111 (N.D. Tex. 1976); Christonsen v. Iow,. 117 F. Supp. 42:1 (N.D. Iowa 1976); Usery v. Meyer Memorial Hosp., 428 F. Supp. 1368, 1372 (note 8) (W1). N.Y. 1!i77).
2 U.S.C. 141 et seq. (1970).
2 Allied Chemical Workers v. Pittsburgh Plate Glass Co., 404 U.S. 157 (Z971). '2 U.S.C. 152(2) (1970).






39

It should be noted that the interpretation of "state or political subdivision thereof" for purposes of the defi-nition of "employer" under the L.M.R.A. is an ongoing process. There exist a number of retirement systems which are commonly thought to be maintained by states or political subdivisions thereof, but which might, if ever challenged, be found to be maintained by an "employer" for L.M.R.A. purposes. Certain kinds of state universities and quasi-public utility and transit authorities, for instance, might be found to be maintained by "employers" for purposes of the L.M.R.A., were the status of certain employers ever questioned.
COMPREHENSIVE EMPLOYMENT AND TRAINING ACT
Public employee retirement systems are affected by the Comprehensive Employment and Training Act of 1973.1 Under the Act, federal grants are made to state and local governments which in turn may employ persons in public service jobs. Such employees must be compensated for wage and fringe benefit purposes at the same levels and to the same extent as other employees of the employer. Additionally, the Emergency Jobs and Unemployment Assistance Act of 1974,3 which amends C.E.T.A., requires that public service woikers under the Act engaged in construction-type jobs must be paid the prevailing wages and fringe benefits, including interests in employee benefit plans, for similar construction in the locality.4 The requirements in these Acts that employees receive fringe benefits comparable to non-Act employees result in the acquisition by public service workers of interests in state and local public employee retirement systems.
BANKRUPTCY ACT
The 1976 revisions to the Bankruptcy Act are of extreme importance to public employee retirement systems in the conceivable event that a local government entity becomes bankrupt. The 1976 revision recognizes that "A municipal unit cannot liquidate its assets to satisfy its creditors totally and finally. Therefore the primary purpose of Chapter IX is to allow the municipal unit to continue operating while it adjusts or refinancies [sic] creditor claims with minimum (and in many cases, no) loss to its creditors." 2 Complex procedural requirements are set out in an attempt to achieve a fair balance between creditors of the bankrupt municipality and the municipality itself, with special consideration given to the special nature of a municipality in terms of its assets and its obligations to its citizens.
When a municipality enters into the procedures contained in the Bankruptcy Act, the retirement systems of that municipality enter a stage even more novel and unsettled than municipal bankruptcy itself. Depending upon the relationship between the pension plan and the bankrupt plan sponsor, the pension plan may be in the position of a creditor to the bankrupt municipality, seeking to force the municipality to honor its unpaid funding obligation to the plan, or it may be in the position of an asset of the bankrupt municipality, eagerly viewed by creditors as a liquid asset capable of satisfying the claims of
29 T.S.C. 801 et seq. (Supp. V 1975) .
2 29 U.S.C. 848(a) (2) and (4) (Supp. V 1975).
3 Pub. L. No. 93-567, 88 Stat. 1845 (1974).
4 2:; U.S.C. 964 (Supp. V 1975).
Pub. L. No. 94-260, 90 Stat. 315 (1976).
2 Rep. . 94-686, 94th Cong., 1st Sess. 6, reprinted [1976] U.S.C.C..&N. 539, 543.






40

deservingz creditors. Conceivably, the retirement system could be both a creditor and an asset of the debtor municipality. Obviously, the prov isions of Chapter IX of the Bankruptcy Act are of extreme significance to a public employee retirement system upon the bankruptcy of the sponsoring municipality.

MILITARY SELECTIVE SERVICE ACT

The Military Selective Service Act 1 affects the operation and provisions of state and local government retirement systems. The Act is intended to insure that veterans will have available to them the jobs they left in order to enter military service. The Act generally requires that a veteran who timely re-applies for his pre-military jobmust be restored to a position of like seniority, status, and pay. The law provides that for purposes of insurance or other benefits offered by the employer, the returning veteran must be considered as having been on furlough or leave of absence during his period of training and service in the armed forces. Certain limitations relating to the quali 'fications, of the veteran, availability of the position, and so on, are found in the statute.
IPrior to 1974, it was clear that the law applied on a mandatory basis to all private employers 2 and to the federal and District of Columbia governments as employers.3 It was also clear that the Act did not apply on a mandatory basis for state and local governments.4 Congress simply suggested that state and local governments as employers afford these re-employment benefits to returning veterans."
In 1974, Congress amended the Act 6 to subject state and local government employers to the same requirements which had been imposed on the federal government and District of Columbia governments in their employer capacity, and on private employers, since the enactment of the predecessor Act in 1940.
Whether an employer must credit a returning veteran with service for vesting purposes, benefit accrual purposes, and other pension and
welfare plan purposes, has frequently been litigated with regard to employers other than state and local governments. There is no reason to believe, however, that the same interpretations of the Act will not be applied to those situations involving state and local government employers made subject to the Act by the 1974 amendments.
Alabama Power Company v. Davis I is a very recent Supreme Court case directly on point. Alabama Power Co. maintained a fairly typical defined benefit pension plan in which the accrued benefits were determined by the number of years of credited service the plan participant had attained. Davis had left the employ of Alabama Power Co. to
enter the military, and he returned to, the Company following his discharge. The Company refused to credit Davis for purposes of benefit accrual for those years in which he was in the Service.
1.38 U.S.C. J 2021 et seq. (Supp. V 1875).
2 0 App. U.S.C. 459(b) (1970).
3 Id.
MeLaughlin v. Retherford 207 Ark. 1094, 184 -S.W, 2d 8 14)
6 50 App. U.S. C. 4.59)(b)(2) (Z) (1970).
1 38 U.S.C. 2021 (SUpp. V 1975).
7431 U.S. 581 (1977).





41

The Supreme Court first takes note of the conflict among the circuits with regard to the issue raised by appellee Davis.8
The High Court then reviews the cases. it had already decided under the Military Selective Service Act. It concludes, that if the benefit at issue
would have accrued, with reasonable certainty, had the veteran been continuously employed by the private employer, and if it is in the nature of a reward for length of service, it is a "perquisite of seniority" [and must be credited to the employee]. If, on the other hand, the veteran's right to the benefit at the time he entered the military was subject to a significant contingency, or if the benefit is in the nature of short term compensation for services rendered, it is not, an aspect of seniority under section 9 [and hence need not be credited].9
The Court proceeds to consider the Alabama Power Company pension plan under this framework and concludes that the accrual of benefits in the instant plan is a perquisite of seniority and accordingly must be granted to Davis.
It is interesting to note those factors which persuade the Court that benefit accruals in the Alabama Power Company plan are more a reward for length of service than in the nature of short term compensation for services rendered. The automatic earning of accruals upon the attainment of a certain amount of service is cited. The Court notes the lengthy (20 years, or 15 years and age 50) vesting schedule of the preEIRJSA plan, and strongly infers from this fact that credit for all purposes under the plan is in the nature of a reward for services. The functionn of pension plans in the employment system "-to reward lengthy service with the same employer-also contributes to the Court's finding.'0
Alabama Power Co. resolves many of the questions raised by the .Military Selective Service Act as it applies to pension and welfare plans, including plans of state and local governments. Service for all purposes in a defined benefit pension plan apparently must be credited to a returning veteran. Presumably vesting schedules somewhat shorter than the one, at issue before the Court will not alter the Court's analysis. One wonders, however, if a greater recognition by the Court that pension benefits are more in the nature of deferred compensation than a reward for lengthy service would alter the outcome. By favorably citing its earlier decisions in Foster v. Dravo Corp." (involving vacation pay) and Accardi v. Pennsylvania Railroad Co.'2 (involving severance pay), and indicating that these holdings are consistent with the rationale expressed in Alabama Power Company, the Court indicates that employers generally must credit military service for severance pay purposes but need not do so for vacation pay purposes-.
Some clear analogies can be draw-n from these examples. Health benefits and prepaid legal plan benefits in most instances seem very much like vacation pay benefits, in that they are intended as a benefit in the nature of short term compensation.
2 Compare Jackson v. Beech Aircraft Corp. 517 F. 2d 1322 (C.A. 10th, 1975) in which it was held no credit need be given for time in military for purposes of pension benefit accruals, longevity pay, vacation beefits or sick leave, with Litwicki v. Pittsburgh Plate Glass Indus., Inc., 505 F. 2d 189 (C.A. 3d 1974), iwhich it was held that no credit need be given under the Act for purposes of vesting or benefit accrual, and Smith v. Industrial Employers and Distributors Ass'n., W4 F. 2d 314 (C.A. 9th.. 1976), in which it was held that credit must be given for purposes of accrual of pension benefits.
Supra, note 7 at 589.
10 Supra, note 7 at 594.
11 420 U.S. 92 (1975).
12 383 U.S. 225 (1966).





42

But much is left unclear. Whether credit for purposes of longevitypay must be credited is unresolved. Similarly, the meaning of the Act in the context of sick pay and disability plans is left unclear. In some, ways these benefits are more like perquisites of seniority than short term compensation. In other ways, the contrary is true. No doubt the. wording of the plan at issue in every instance will determine to a great extent the nature of the benefit. Whether the Supreme Court's analysis. will change if the plan is a contributory one is obviously of importance to governmental plans.
Also of importance is the Supreme Court's express reservation regarding whether defined contribution plans are to be treated differently from defined benefit plans for purposes of the Act.13 A defined contribution plan was not at issue in Alabama Power Company, and thus no. direct inferences can be drawn from the Court's reservation on this issue. Given the large number of defined contribution plans in the public sector, particularly among the retirement systems of colleges and universities owned by state and local governments, the resolution of, this issue is of some significance.
In summary, the Military Selective Service Act, made mandatory in terms of state and local government employers in 1974, represents another federal statute presently affecting public employee retirement systems. As greater numbers of returning veterans in future years. resume employment with state and local government employers, the Act and its interpretations will increasingly need to be considered by public employee retirement systems.
COST ACCOUNTING STANDARDS BOARD
Legislation establishing the Cost Accounting Standards. Board and regulations issued by the Board,2 affect those few public employeeretirement systems which cover employees of defense contractors and subcontractors. State universities periodically enter into defensecontracts to conduct technical research and development, and the pension component of such contractor's cost must be computed and measured under the regulations issued by the Cost AccountingStandards Board. To the extent a public plan covers employees of contractors subject to the C.A.S.B. regulations and accounts for thecost of that coverage in a manner inconsistent with the C.A.S.B.. requirements, this aspect of existing federal regulation of. public: employee retirement systems is meaningful.
13 Supra, note 7 at 593 (footnote 18).
1.50 App. U.S.e. 2168 (1970).
2 4 C.F.R. 331.30; 4 C.F.R. 412.10-412.80 (1977);











PART HII-STATE LAWS APPLICABLE TO PUBLIC
EMPLOYEE RETIREMENT SYSTEMS
Of obvious significance to public employee retirement systems are, the constitutional, statutory, and common law provisions of the state in which the governmental plan is located. A listing of many of these, laws is contained in Appendix V of this Report.
The variety, in both scope and nature, of state law provisions. relating to public employee retirement systems, is extensive. -Many states have constitutional provisions which prohibit the impai rment of contracts and require the state to afford clue process before dlepriving persons of life, liberty or property. Numerous states, by constitu-. tional or statutory provision, characterize plan interests as contractual in nature. Plan provisions relating to benefit levels, contribution. formulae, eligibility, and funding requirements are frequently addressed at length by statute. Some states spell out the nature of investments that must be made with retirement fund assets. Fre-. quently, common law is used to address an issue that has not been, directly addressed by the legislature.
In many states, no constitutional or statutory provisions are found, in vital areas such as reporting and disclosure to participants, review
-of claims procedures, auditing and accounting standards, and fiduciary responsibility. No state appears to have provisions relating to the, insuring of unfunded plan liabilities upon the termination of the plan.
The sheer complexity of the state regulatory framework is in itself significant. Every state except Hiawaii has more than one pension system. Forty states have ten or more plans. The number and variety of state laws affecting each plan has unquestionably produced confusion among plan participants. The problem is compounded by the frequ ent absence of uniform or enforceable reporting and disclosure requirements. Some public employee retirement systems have, how-. ever, adopted beneficial plan practices in the absence of a legal requirement to do so. In the reporting area, for instance, many plains prepare and distribute summaries of major plan provisions, and periodically report publicly on the financial condition and related m atters of the plan. Some states and localities have established a, claims review procedure to enable a participant whose claim for benefits has been denied an opportunity to obtain a more independent review of his application. The absence of any legal mandate for these beneficial practices substantially lessens their value to plan participants. Public plan practices that are not required by any statutory orconstitutional provision can be eliminated as easily as they are established. F requently, a participant may not have standing to -enforce, plan practices that have been established only informally. As a general matter, plan participants receive greater protection from beneficial plan practices that are mandated by law rather than voluntarily introduced.
(43)

74-365-T 8-





44

A closely related matter involves the interpretations which state ,courts have given to state statutory and constitutional provisions relating to governmental benefit plans.
In some instances, state law has been interpreted in a manner which has served to strengthen the protection afforded plan participants.
The Washington Supreme Court, for instance, in Bakenhu's v. C ity of Seattle,' finds under state law that a pension is deferred compensation for services rendered, and hence contractual in nature.2 The Court goes on to state that a plan participant, upon satisfaction of the prescribed conditions, is entitled to receive the plan benefit for which he contracted, andi that pension rights "may be modified prior to retirement, but only for the purpose of keeping the pension system flexible and maintaining its Integrity." I When confronted with legislation which purported to retroactively reduce the plan participants' accrued benefit, the Court acts upon its contract-like analysis and strikes down the legislation.
A similar example of a state law interpretation which serves to benefit plan participants is found in Opinionu of the Justices.4 At issue was a proposed statutory enactment which would raise the mandatory employee contribution rate from 5 percent to 7 percent of salary, with no increase in benefit levels. The Massachusetts Supreme Court first analyzes relevant state law5 'and concludes that participation in the pension plan establishes a contractual relationship which protects the plan participant "in the core of his reasonable expectations, but not' against subtractions which, although possibly exceeding the trivial, can claim certain practical justifications. . In considering the pro-posed increase in the contribution rate from 5 percent to 7 percent, the Court enhances the protection participants in theory enjoy from the contractual nature of the pension -interest, and strikes down the proposed contribution rate increase as presumptively violative of both state statutory and constitutional as well as federal constitutional provisions.
In Sgaglione v. LeVitt,7 the New York Court of Appeals is faced with a New York statute which purported to compel the trustee of certain state pension funds to make specified investments with pension fund assets, thereby divesting the trustee of his discretion with regard to the investment of plan assets. The Court takes note of the nonimpairment clause of the New York Constitution,8 and concludes that:
To strip this person [trustee], in this instance the State Comptroller, an independently elected official it so happens, of his personal responsibility and commitment to his oath of office, is to remove a safeguard integral to the scheme of maintaining the security of the sources of benefits for over a half century ... But it is . concluded that the Legislature is powerless in the face of the constitutional nonimpairment clause to mandate that he mindlessly invest in whatever securities they direct. ..
The statute the New York Court struck down was a cornerstone of the New York City MAC financing program of 1975, and a crisis
0
1 48 Wash. 2d 695, 296 P. 2d 536 (1956).
1 Id. at 6A8, 296 P. 2d at .538.
3 Id. at 701, 296 P. 2d at 540.
&'3r4 Mass. 847, 30.3 N.E. 2d 320 (1973).
& Mass. Gen. Laws Ann., ch. 32.
*Supra, note 4, at 303 N. E. 2d at M2.
37 N.Y. 2d 507, 337 N. E. 2d 592 (1975).
*N.Y. CONST. art. V, 7.
8Su pra, note 7 at 512-13, 337 N.E. 2d at 595.






45

situation was generally thought to exist. As such, Sgaglione v. Levitt stands as a clear example of a state constitutional -provision which was interpreted to provide a significant level of -protection for govern-mental plan participants.
Act 293, originally enacted by the Pennsylvania legislature in 1972,1 stands as an example of a state reporting statute that is de.Signed to permit strict enforcement. The Act requires all municipal pension systems to retain an actuary for the purpose of periodically preparing actuarial statements and reports to be filed with the State Department of Community Affairs. Importantly, the Act contains a .number of enforcement mechanisms. Failure by a municipality to file such reports results in the withhlolding from such municipality of any
adall state contributions to the delinquent -municipalities' pension funds."' The study may also be performed by the state, and the delinquent municipality in that instance must reimburse the state for the ,Cost. 12
Act 293 is subject to criticism in that it addresses only actuarial ,aspects of the plan. Its substantial enforcement mechanisms, however, serve to make Act 293 a good example of a state reporting requirement which is designed to achieve the purpose for which it was presumably enacted.
It must also be noted that in numerous instances state laws which
-appear to afford significant protections to pension plan participants .have been interpreted in a manner that demonstrates that protection to be, in reality, quite limited.
In re. Enrolled Senate Bill 126.9 '3 stands in direct contrast to Opinion of the Justices.' The Michigan legislature sought an advisory opinion regarding a proposed statutory revision of the public school employees' retirement system. The statute, for a particular class of employees, would require an increase, from 3 percent to 5 percent, in the rate of employee contributions to the plan, with no increase whatsoever in benefits. The Michigan Supreme Court first notes a Michigan .-constitutional provision which explicitly states that accrued governmental pension benefits shall represent a contractual obligation which :Shall not be impaired or diminished.5 The Court then states:
Under this constitutional limitation the legislature cannot diminish or impair .accrued financial benefits, but we think it may properly attach new conditions for earning financial benefits which have not yet accrued. Even though compliance with the new conditions may be necessary in order to obtain the financial benefits which have accrued, we would not regard this as a diminishment or impairment of such accrued benefits. ---15
Thus the Michigan Supreme Court "interprets" the noniinpairment .clause of the State constitution as permitting an increase in the employee contribution rate, precisely the kind of impairment that the Massachusetts Supreme Court found unconstitutional in Opinion of the Justices.
A second example of a state law "protection" that upon testing
-was interpreted as not very protective of plan participants is found in
-People ex rel. Illinois Federation of Teachers v. Lindberg.17 A-n Illinois
10 53 Pa. Stat. Ann. J 730.1 et seq.
11 53 Pa. Stat. Ann. 730.4.
12 Id.
13 389 Mich. 659, 209 N.W. 2d 200 (1973).
14 Supra, note 4.
15Mich. OONST. art. 9, 24.
is Supra, note 13, at 663-64, 209 N.W. 2d at 202--0&
17 60 III. 2d 266, 326 N.E. 2d 749 (1975); cert. denied, 423 U.S. 839 (1975).






46

statute spells out the level of contributions which the state is to make to each of a number of 'State retirement systems.'" The Illinois legrislature enacted a series of appropriations measures designed to bring the state's contributions to these retirement systems nearer the level required by the statute.
It should also be noted that a provision of the Illinois Constitution states:
Membership in any pension or retirement system of the State, any unit of local government or school district, or any agency or instrumentality thereof, shall be an enforceable contractual relationship, the benefits of which shall not be diminished or impaired.19
When the governor vetoed or reduced the appropriations bills designed to more adequately fund the retirement systems at issue, suit was brought in which it was alleged that the veto and reductions by the governor violated the impairment of contracts provision of the Illinois Constitution and the statutory mandate that the state fund the pension plans at the specified level.
The Illinois Supreme Court first rejects the argument that the governor's actions violate the state constitutional provision. Whatever the nature of those contractual rights, the Court indicates, it cannot be "argued that the provision was intended to res-trict the Governor's constitutional authority to reduce or veto a pension appropriation measure."1 20 The Court then holds that the statutory material relating to the retirement systems nowhere establishes a contractual relationship between plan participants and the state orthe plan 21 Finally, the effect of the statutory funding "mandate" in terms of the governor's power to reduce or veto appropriations bills is addressed, and the Court concludes that the funding provisions in no way limit the governor's authority.22
Thus, what appear to be state constitutional and statutory provisions that assure that a plan will be funded at a specified level, upon judicial inteirpretation, are found not to assure any such protection at all.
In conclusion, the benefits and protections that state laws appearto provide plan participants iD Some instances are interpreted in a manner that accomplishes such protections, but in many instances are found to afford none of the benefits or protections that the state statutory or constitutional provision at first glance appears to assure..
11 Ii. Rev. Stat., ch. 10O' (1973).
1g 111. CONST., art. Xl [I, 15.
20 Supra, note 17 at 272,.326 N.E. 2d at 752.
21 Supra, note 17 at 275, 326 N.E. 2d at 753.
22 Supra, note 17 at 277, 326 N.E. 2d at 754-55.












PART IV-CHARACTERISTICS AND 'OPERATIONS OF
PUBLIC EMPLOYEE RETIREMENT SYSTEMS

CHAPTER A-INTRODUCTION
During the course of legislative activity leading to the September 2, 1974 enactment of ERISA (The Employee Retirement Income Security Act of 1974-Public Law 93-406), the Congress amended early legislative versions of the Act which originally covered governmental pension pl ans to exclude such plans from provisions relating
-to reporting, disclosure, fiduciary, and vesting standards, although ,certain Internal Revenue Code provisions were extended to public pension plans by ERISA. At that time persuasive arguments were madle that the problems encountered by governmental plans were unique and deserving of additional study. As a result, Section 3031 was added to ERISA:
The Committee on Education and Labor and the Committee
on Ways and Means of the House of Representatives and the Committee on Finance and thie Committee on Labor and Public Welfare of the Senate shall study retirement plans established and maintained or financed (directly or indirectly) by the Government of the United States, by any State (including the District of Columbia) or political subdivision thereof, or by any agency or instrunentali ty of any of the foregoing. Such study shall include
an analysis of(1) the adequacy of existing levels of participation, vesting and financing arrangements,
(2) existing fiduciary standards, and
(3) the necessity for Federal legislation and standards with
respect to such plans. In determining whether any such plan is adequately financed, each committee shall consider the necessity for minimum funding standards, as well as the
taxing power of the government maintaining the plan.

EXTENSIVE STUDIES MADE'
Ex-tensive research, investigations, and hearings were conducted by thie Subcommittee on Labor Stand ards-Pension Task Force in dischairging this responsibility. Public hearings were held during 1975 in California, Connecticut, Illinois, and Washington, D.C. These hearingrs utilized H.R. 9155, a bill extending all of the Title I and IV standards of ERISA to non-federal governmental plans, as a vehicle to focus attention on the specific implications of legislated federal sta~ndards.
In the fall of 1975 a census of federal, state, and local government retirement systems was initiated in an attempt to identify for the first time the universe of public pension plans. This effort resulted in the identification of the size and major characteristics of 68 federal
(47)






48

plans and 6,630 state and local government pension plans. The characteristics of the plans included in this inventory were first reported in the March. 1 .1976 Interim Report of Activities of the Pension Task Force of tle Subcommittee on Labor Standards (see Appendix IV for a coinplete tabulation of plans by state).
The 22 page questionnaire, "Pension Task Force Survey of Public Employee Retire~ment Systems" (see Appendix 11), was sent to a sample of the PE1_RS universe in order to gather information on public pension plan administration, benefit structure, finances, and funding. Between May and September 1976, survey data was obtained on the pension plans covering 100 percent of the federal government em-, ployees and 96 percent of the state and local government employees (Appendix I displays the tabular results of the survey data for 1975).
In order to obtain additional information on the investment opera-tions of state and local government pension funds, the Subcommittee on Labor Stand ards-Pension Task Force in conjunction with the Pension Research Council of the University of Pennsylvania con-ducted a survey focusing on pension fund investments. Nearly twothirds of the public pension funds with assets in -excess of $50 million, provided information on statutes,- policies, and practices affectingpension fund investments in 1974-75.
In addition, extensive research was undertaken to record and analyze the federal, state and local laws affecting public employee retirement. systems (see Parts 11 and III of this report and Appendices V through XI).
PERS INFLUENCt VAST AND DIVERSITY GREAT
The various studies carried out by the Pension Task Force thorough-ly document the nature and scope of the characteristics and operations of public employee retirement systems. In their entirety, public employee retirement systems (PERS) exert a substantial influence on' the economic, social, and political fabric of the United States.
The national securities and other capital formation markets are, heavily influenced by state and local government pension fund in-vestments exceeding $108 billion (1975). Such investments have recently been increasing by 13 percent annually. State and local pension, plan benefit payments ($7.3 billion in 1975) have been increasing even more rapidly at a rate of 16 percent annually. Federal, state, and local pension benefit payments add over $21 billion annually to the consumption and savings elements of the economy. Federal, state,, andl local governments must provide for tax revenue exceeding $23 billion annually in order to meet employer obligations to public: employee pension plans.
Rhe strengths, weaknesses, and diversity of public employee retirement systems are discussed in chapters B through H of this report. Chapter B discusses some of the general characteristics of the 6,698 federal, state, and local retirement systems (including their relationship to Social Security) which cover 12.7 million state and local planI participants and 9.5 million federal plan participants. Chapter C discusses the serious deficiencies which exist among public pension plans in the reporting, disclosure, accounting, auditing, recordkeeping, and other administrative aspects of plan operations.






49

Chapter D discusses publIic pension plan benefit protections in relation to ERISA's provisions in the areas of eligibility, vesting, portability, and plan, termination insurance. Chapter E describes the benefit structure of public pension plans and the extent to which the combined benefits from public pension plans and Social Security replace the pre-retirement income of retired public, employees.
Chapter F provides informa-tion on the finances of public p'ensio n plans and the restrictions which impair the stability and pr'edictability of employer pension contributions. Chapter G documents the serious funding problems of many public pension plans and the lack of adequate actuarial valuations, measures, and standards.
In discussing the fiduciary and investment practices of state and local government pension funds, Chapter H points out that the presence of various investment restrictions, the absence of adequate disclosure and other safeguards, and the lack of fiduciary accountability all contribute to a high potential for fiduciary -abuse and the. loss of pension plan assets and income.
In consideration of the fundamental national interests involved and the substantial body of federal law and federal involvement. in governmental pension plans, the facts presented in this report show a compelling need for a revised and expanded set of federal standards applicable to federal, state, and local pension plans. Serious consequences are likely unless immediate steps are taken by the federal, state, and local governments to reiedy the serious deficiencies that exist among, the various retirement systems for public employees.

TECHNICAL NOTE
The reader should be careful to distinguish between the PERS, universe of plans given in Appendix IV and the tabular data from the Pension Task Force Survey of a sample of the plan universe as given in Appendix 1. Appendix III provides an explanation of the survey methodology. While it was intended that the study include only those governmentall plans" presently exempted from ERISA (Public Law 93-406), the inclusion of a particular plan in this report or in the survey should not be taken as proof of such statutory exclusion. When clarifying regulations are issued, it may in fact be the case that the same plans now maintaining their exempt status under ERISA, particularly those of a "quasi-governmental" nature, may be found to be covered plans.
The data from the Pension Task Force Survey of Public Employee, Retirement Systems in Appendix I is presented in an extremely detailed format in order that the information be of maximum usefulness to plan administrators, public officials, public employees, and others having an interest in public employee retirement systems.
The data items in the tables of Chapter B through HI may not add to exactly 100.0 percent due to rounding.

ACKNOWLEDGEMENT
This study was a mammoth undertaking and could not have been, accomplished but for the willing cooperation of the many persons who, spent considerable time and effort in supplying the basic survey





50

information. The assistance of the many state and local pension plan administrators, PERS organizations, and public employee organizations too numerous to mention individually is (gratefully acknowledged.
Special acknowledgement is due to Vance Kane, Chief of the Financial Branch-Government Division, Bureau of the Census, who provided the Pension Task Force with the initial data base from which the expanded universe of state and local pension plans was constructed, and to the following persons in the Division of Financial and General Management Studies, U.S. General Accounting Office: Erwin Bedarf who assisted in developing and testing the Pension Task Force Survey questionnaire, Herbert R. Martinson and his staff, who were responsible for the questionnaire sampling and editing process, and Norman Daley, who produced the required survey tabulations.
Special thanks is extended to Raymond Schmitt of the Education and Public Welfare Division, Congressional Research Service, who made important contributions to all phases of the study, and to Howard Zaritsky of the American Law Division, Congressional Research Service, who supplied helpful legal assistance and analysis. Appreciation is also expressed to John Dean and Suzanne Hays of the Pension Task Force for their support in the preparation of this report.
S. HOWARD KLINE,
Counsel and Staff Director.
RUSSELL J. MUELLER,
Actuary and Minority Legislative Associate.












CHfAPTER B-PERS UNIVERSE-GENERAL CHARACTERISTICS
The programs making up the public employee retirement system (PERS) have been found to be as varied as the pension plans which make up the private pension system. In 1975 over 6,698 federal, state and local government retirement systems were identified as covering 15.4 million civilian and military employees. In addition to the 6,698 plans, there are other arrangements made by sponsoring governments to provide retirement income for their employees. These programs include deferred compensation contracts, salary reduction plans, and "tax sheltered" annuities. According to the Institute of Life Insurance, .750,000 individuals were covered and 20,000 persons were receiving benefits in 1974 under 403 (b) tax sheltered annuity plans. It i's believed that, at. least a majority of the 750,000 persons covered under such plans were public school teachers and other public employees--. Trhe Employee Retirement Income Security Act of 1974 established another type of tax deferred pension plan, the Individual Retirement Account and the Individual Retirement Annuity (IRA), which was available'for the first time in 1975 to all employees working in the public as well as the private sectors -who were not covered under any other public or private pension plan. It remains to be seen to what .extent pension coverage will be expanded through the use of IRAs to the approximately 1.5 million public employees not presently covered under any public plan.
NUMBER AND MEMBERSHIP
The findings of the Pension Task Force as to the total number and membership of the PERS are summarized in Table B1. In 1975, 6,630 pension plans, of one type or another were identified as being 'Maintamned by state and local governments for about 10.4 million full- and part-time workers. Another 2.3 million persons were receiving retirement, disability, or survivors benefits under such plans or were otherwvise eligible to receive benefits at a later date. TABLE B1.-NUMBER AND MEMBERSHIP OF PUBLIC EMPLOYEE RETIREMENT SYSTEMS OF FEDERAL, STATE, AND LOCAL GOVERNMENTS, 1975
Membership (thousands)
Number of
Level of government, plans Active Nonactive Tota I
State and local----------------------------------- 6, 630 10, 387 2 347 12, 744
Federal (uniformed services)-------------------------- 4 2, 181 1,1 094 3, 275
Federal (nonuniformed services)----------------------- 64 2, 839 3, 402 6, 241
Total------------------------------ ---- 6,698 15,417 6,843 22,260
Note: See app. IV for more detail
(51)








052











T, jm gig r 2;; silk-! NPI:Ao _- -0a 1%'*ftm:! 2
0 A to
ri

F181,42 SIA22,40 Co IV 0 40 a

:IRAs T -I g1k
A Ck F%




LU
cc
Cg's W,
cz . . . . .
uj .0
cr z Z
X: u
1*e 0- ar
C ef 0 -W fui a
< E I14
U.: v 0 g 1 13 i :
0 w t 0 = r.
c-S c .2 E E S I tz
Rc ;S V4 c -2 u C, -2 ; X j i a (% :
Z ; a ,
r3,% 'S M a x j: e -4 IA-Z -j gv
E,2 i tA b 7E t,414
< .,I
:j Jr- i 78S 4 Z tl
A : Zn v u w 4L 46
Wa
Wc 94 1














CD


ca

CS
go Cc



IN







CD cc









Cc
CN 9%
Wo a) W6

C


a) C*4



C3
LE WTAmF, B2






53

In 1975, the federal government, including its agencies and instrumentalities, maintained 68 employee pension plans, the largest ones beig the Military Retirement System (2.1 million active members) and the Civil Service Retirement System (2.7 million active members). The remaining 66 systems have about 183,000 active members or about 5.6 percent of the total active membership in all federal plans. The roughly 4.5 milion persons presently receiving or ,expecting to receive benefits under all 68 federal systems represent 89 percent of the total federal active membership. By way of contrast, the corresponding inactive to active membership ratio for all state :and local systems is 23 percent. About half of this marked difference is due to the fact that terminated employees with vested benefits Make up 45 percent of all federal inactives while the corresponding figure for state and local plans is 11 percent. Some of the causes for this variation in the number of terminated vested employees are discussed in the next Chapter. The ratio of inactive to active members for the federal system is over tvice that for the state and local system,
-even ignoring terminated vested employees. This demonstrates the fact that the state and local part of the PERS as a whole has yet to reach the degree of maturity attained by the federal system. In this regard the state and local system is more akin to the private pension :system (see Table G12).
The number of state and local plans by size of active membership is shown in Table B3. One striking similarity with the private pension
-system is the large percentage, nearly 80 percent, of all plans with fewer than 100 active members. At the other extreme are the 390 plans with 1,000 or more active members. While making up less than 6 percent of the total plans, these plans cover about 95 percent of the active membership in all state and local government pension plans.
TABLE B3.-NUMBER OF STATE AND LOCAL PUBLIC EMPLOYEE RETIREMENT SYSTEMS BY SIZE, 1975
Number of Percentage
plns of tot
.Size of active membership:
0 to 4 --------- ------------------------------------------------- 793 13.7
2to 1 ...-------------------------------------------------------1,545 26..7
25 to 99 --------------------------------------------------- 1,110 19.2
W, to 199. ------------------------------------------------------ 332 $*
:200 to 99.17 3
500 to 999 ------------------------------------------------------------ 187 k
1,00ato 4,999 --------------------------------------------------------- 2Q6 3.6
5,060 to 9,94--------------------------------------------------------- 6 1.0
1%9ag ove r 2Uknown-.- --- ..1,14
T ..taL------------------------------------------------------- 5,7 7 0.O,
Total number of plant exdude 842 plans for which data by type of plan and size of membership. is unavailable. lrthaerstudy reveled that the Plans for which active memlbership i; unknown are princ4pHy Local pl ns for policem
-and lrmen having fewer than 100 active members.
DEFINED BENEFIT AND DEFINED CONTRIBUTION
Among the 6,698 retirement systems are defined contribution plans as well as the mre typical defined benefit or benefit "formula" plans. A participants benefits under a defined contribution plan are based solely on the amotints ooutributed to the participant's account and .any income earned thereon. The largest number of similar defied contribution plans are the 314 plans, maintained by cities and counties for their highly mobile managers and administrators, These plans ,are administered by the ICMA Retirement Corporation. At the end





54

of 1974, 605 employees participated in the 314 plans. The next largest group of similar defined contribution plans are those maintained by institutions of higher education for faculty and other personnel. In 1975, 294 such plans were in operation (see Table B4) and all the plans were funded through Teacher Insurance and Annuity Association of America and its affiliate, the College Retirement Equities Fund (TIAA-CREF). In addition to the above, public employees also belong to a relatively small number of defined contribution plans most of which are insured and of the money-purchase type.
While only about 2.2 percent of state and local employees are members of defined contribution plans, nearly 16 percent are covered under "combination" plans having both defined contribution and defined benefit features. The bulk of all covered employees, 82 percent, are members of defined benefit plans providing the more traditional pay-related and formula calculated benefits. Together, defined benefit and combination plans comprise over 85 percent of all state and local plans. With the exception of several small thrift plans and TIAACREF plans, all federal plans are of the defined benefit type. Chapter E presents extensive information on the varied benefit structures of both defined benefit and defined contribution plans.

INSURED AND NON-INSURED
The detailed distributions shown in Table 35 of Appendix I reveal that 27.7 percent of all state and local plans provide some or all retirement benefits through insurance companies (TIAA-CREF is
included in this category). The plans so insured tend to be small, however, and cover only 4.8 percent of all active employees.

CONTRIBUTORY AND NONCONTRIBUTORY
About 75 percent of the state and local plans surveyed require employees to make contributions. Most of the plans now "closed" to new members are or were at one time contributory plans. In total about 85 percent of all active state and local employees are required to contribute to their plans. In contrast, about 44 percent of the federal plans covering 56 percent of all employees require mandatorv contributions.
VWhile the plans for general government employees at the federal, state, and local level share many common characteristics, the plans covering the uniformed employees differ in several important respects. The federal military plans are wholly noncontributory while the plans at the state and local level covering uniformed employees (police and fire) are nearly all contributory. This difference may be attributable to the fact that all members of the federal uniformed services contribute to Social Security while two-thirds of the state and local uniformed employees do not.
A small but growing number of state and local plans are reducing or eliminating mandatory employee contributions. Presently, state employees in Florida, Michigan, Missouri, and New York are included in the noncontributory category, totaling 21.4 percent of all state plans and 15 percent of all state employees. For some employees in Now York City and in Wisconsin, governmental employers have "picked up" a portion of the contributions which would otherwise have been made by the employees.






55

A few plans give employees the Option to either make voluntary contributions in order to receive retirement benefits or to waive pension coverage. At the federal level the Vice President, Senators, Representatives, and congesoa employees are given the option to contribute to the Civil Service Retirement System. At the local level many plans for volunteer firemen operate on a voluntary basis.
Among the category of contributory plans are 23 or more plans which provide retirement benefits "supplemental" to basic benefits paid from other plans maintained at the state or local level. Several universities maintain supplemental plans which are of the noncontributo :',y pay-as-you-go type.

PLANS BY JURISDICTION
Over two-thirds of all state and local government retirement systems are found in the 10 states with the largest number of plans (see Table B4). With over 1,413 plans, Pennsylvania alone has over one-fifth of the total plans. Among the top five states and following Pennsylvania in order are Minnesota (638 plans), Illinois (465 plans), Oklahoma (435 plans), and Texas (398 plans). By way of contrast to the decentralized, nature of the systems in the majority of states is the single statewide plan in Hawaii covering all public employees in that state. While the states with the smallest populations tend to have the least number of plans, this is not entirely the case. Other states such as New. Mexico (with 4 plans), Ohio (with 7 plans), and Oregon withh 9 plans) have achieved major gains in bringing all public employees in their respective states under a small manageable number of systems.
TABLE B4.-Number of Federal, State, and local public employee retirement systems by State or other jurisdiction, 1975
Number Number
State or jurisdiction: of pians 1 State or jurisdiction-Con, of Plans 1
(1) Alabama---------47 (31) New Jersey------------39
(2) Aak---------8 (32) New Mexico------------4
(3) Arizona----------------10 (33) New York -------------117
(4) Arkansas --------------69 (34) North Carolina ---------58
(5) California--------------72 (35) North Dakota-------21
(6) Colorado--------------343 (36) Oi----------7
(7) Connecticut ------------105 (37) Oklahoma -------------435
(8) Delaware--------------13 (38) Oregon-----------------9
(9) District of Columbia-- 7 (39) Pennsylvania --------- 1, 413
(10) Florida ----------------335 (40) Rhode Island ----------22
(11) Georgia---------------54 (41) South Carolina---------13
(12) Hawaii -----------------1 (42) South Dakota-----------7
(13) Idh---------11 (43) Tennessee--------------27
(14) Ilni--------465 (44) Texas-----------------398
(15) Indiana---------------249 (45) Uth---------12
(16) Ioa---------75 (46) Vermont---------------31
(17) K ansas ----------------55 (47) Virginia ---------------28
(18) Kentucky--------------49 (48) Washington- - ----------5 3
(19) Louisiana---------62 (49) West Virginia ----------69
(20) Maine -----------------7 (50) Wisconsin--------------46
(21)' Maryland--------------39 (51) Wyoming ---------------8
(22) Massachusetts-_---103 (52) Puerto Rico-------------(23) Michigan-____-----187 (53) Virgin Islands -----------1
(24) Minnesota -------------638 (54) Guam ------------------1
(25) Mississippi--------23 (55) Washington Metropoli(26) Missouri---"7------------52 tan Area Transit Au(27) Montana--------------28 thority---------------1
(28) Nebraska --------------52 (56) Fdrl--------68
(29) Nevada----------_-----10 __(30) New Hampshire 9 Total--------------6,(159
2 Total number of plans exclude WO9 plans for which data by state Is unavailable.






56

Pension plans covering every imaginable description of public employment are found at all governmental levels. In size, they range
from the 417,000 member New York State Employees' Retirement
System, which is the largest system at the state and local level, to the Evergreen Park, Illinois Firemen's Pension Fund, which at the time the census was taken did not have any active members. The Washington Metropolitan Area Transit Authority (better known as
METRO) was found to be a "body corporate and politic" and its unique 1,000 member system is included in the "special district"'
category (52) as shown in Table B5. Some other systems which cover unique categories of workers are the Central Intelligence Agecy Employees Voluntary Investment Program, the TIAA-CREF Retirement Plan for Faculty Members of the Uniformed Services
University of the Health Sciences, the Waterford (California) Irrigation District Pension Plan, the Dekalb County, Alabama Hospital Annuity and Benefit Plan, the Salt River Project Agricul-. tural Improvement Power District Retirement Plan (Arizona),
and the State of South Dakota (ement Plant Pension Plan.

ADMINISTRATION AND COVERAGE CATEGORIES
Generally, public employee retirement systems may be categorized by level of administration-that is, at the federal, state, city, county, township (borrough), or special district level. As shown in Table B5, over 80 percent of the plans are administered at the city and township levels, while the state governments administer 9.6 percent of the total plans. Plans covering either policemen or firemen exclusively make up two-thirds of all plans. While 68 plans have been categorized as
being "teacher plans," teachers in 21 states participate in a statewide system covering state employees as well as teachers. In most states judges are covered under separate retirement plans while legislators in seven states have separate systems.
TABLE B5.-PUBLIC EMPLOYEE RETIREMENT SYSTEMS OF STATE AND LOCAL GOVERNMENTS BY TYPE OF ADIMINISTERING GOVERNMENT, 1975
Number
Level of administration of plans I Percent
STATEWIDE PLANS
1. State employees only ------------------------------------------------------ 30 5.4
2. State enplyees plus local government erTnlyees (knluing teachers) -------------- 43 7.8
3. Police only ----------------------------------------------------------------- 31 5.6
4. Fire only ---------------------------------------------------------- 13 2.3
5. Police and fire only --------------------------------------------------- 8 1.4
6. Teachers --------------------------------------------------------- 35 6.3
7. Legislators ----------------------------------------------------------------- 8 1.4
8. Judges ----------------------------------------------------------- 41 7.4
9. Local government employees (excluding teachers) ------------------------------- 22 4.3
10. Local government employees (including teachers) ------------------------------- 0 .0
11. Faculty and others (TIAA-CREF only) -------------------------------------- 294 53. 1
12. Faculty and others (other than TIAA-CREF) ----------.------------------------- 13 2.3
13. All other statewide plans ---------------------------------------------------- 16 2. 9
Total (9.6 percent of all plan) ----------------.-------------------------- 554 100. 0
CITY PLANS
21. City employees only (excluding teachers) -------------------------------------- 445 12.9,
22. City employees only (including teachers) ------------------------------------ 5 .1
23. Police only ---------------------------------------------------------------- 940 27.3
2!. Fire only ----------------------------------------------------------------- 1,710 49. 7
25. Police and fire only --------------------------------------------------------. 276 8. 0
26. beaches --------------------------------------- ------- .---------------- 23 .7
27. Judges ---------------------------------------------------------------- 2 .1
28. All other city-administeied plans- ..-----------.--.--------------------------- 40 1.2
Total (59.5 percent of all plans) ----------.------------------------------ 3, 441 100.0
See footnotes at end of table.







57

TABLE 85.-PUBLIC EMPLOYEE RETIREMENT SYSTEMS OF STATE AND LOCAL GOVERNMENTS BY TYPE OF' ADMINISTERING GOVERNMENT, 1975-Continuedl

Number
Level of administration of plansl Percent.

,C0ttTY -PLANS
31. County employees only (excluding teachers) ----------------------------------- 181 6.
32. County employees only (including teachers)------------------------------------- 3 1. 1
33. Police only ------------------------------------------------------------ 29 10.8
34. fire only-------------------------------------------------------------- 22 8.2
35. Police and fire only ------------------- ------- -- -- -- 7-- -- -- -- -- -- -- -- -- -- ---3 1.1L
36. Teachers-------------------------------------------------------------- 7 2.6
37. Judges------------------------------------------------------------------7.138. All other county-administered plans - --------------------1 .
Total (4.6 percent of all plans)------------------------------------------- 268 100.0
TOWNSHIP PLANS
41. Township employees only (excluding teachers) --------------------------------- 473 36. 2
42. Police only ------------------------------------------------------------ 772 59..1
43. Fire only-------------------------------------------------------------- 34 2.644. Police and fire only ------------------------------------------------------ 9 .7
45. All other township-administered plans ---------------------------------------- 19 1. 4
Total (22.6 percent of all plans) ----------------------------------------- 1, 307 100. 0
SPECIAL DISTRICT AND OTHER PLANS
51. Teachers-------------------------------------------------------------- 3 1.4
52. All other (e.g., transit authofiies)------------------------------------------ 215 98.6
Tbt~l (3.8 percent of all plans)------------------------------------------- 218 100.0
ALL PLANS
Police (3)+(23)(33)+(42)------------------------------------------------- 1,772 30.6
Fire (4)+(24)+(34)+(43)--------------------------------------------------- 1,779 30.7
Police and fire (5)+(25)+(35)+(44)------------------------------------------- 295 5. 1
Teachers (6)+(26)+(36)+(51) ----------------------------------------------- 68 1. Z
Faculty and others (11)+(12) ------------------------------------------------ 307 5. 3
Legislators (7)------------------------------------------------------------- 8 .1
Judges (8)+(27)+(37) ---------- ------------------------------------------- 47 8
State (1)+(2)+(13) -------------------------------------------------------- 89 1.5
Local (9)+(10)+(21)+(22)+(28)+(31)+(32)+(38)+(41)+(45) --------------------- 1, 208 20. 9
Other (52) --------------------------------------------------------------- 215 3.7
Total State, and local government plans ----------------------------------- 5,788 100.0

1Total number of plans excludes 842 plans for which data by level of administration is unavailable.

While police and fire plans constitute nearly two-thirds of the PERS plan universe, Table B6 shows that the employees in the police and fire category make. up only 6.7 percent 'of all covered employees. The teacher category is the largest with 30 percent of all covered employees. State employees make up 22.8. percent, and local government employees make up 26.7 percent of the total coverage. Employees of transportation, utility, and other quasi-governmental organizations form a significant 9.7 percent of the total number of covered workers.
Although the split between the number of, full-time and part-time employees '.was unavailable for many of the -plans surveyed (see
Appendix I, Table 14), it remains that a significant proportion of covered part-time workers are volunteer firemen (Table B6 shows 24.1 percent to be in this category). The teacher category makes up 26.9 percent of the total part-time coverage, while the state and the local. employee categories constitute 1.7 percent and 8.8 percent respectively. 'The' large -percentage difference between the teacher category .and the state and local categories is 'probably due to the
larger absolute number of part-time teachers, the less restrictive plan eligibility requirements for part-time teachers, and perhaps a greater understatement of the number of part-time employees in the state and local categories.






58
TABLE B6.-DISTRI BUTTON OF RETIREMENT SYSTEM ACTIVE EMPLOYEES BY EMPLOYMENT CATEGORY
Percent of employees
Less than
Full-time full-time
Employee category employees employees Total
State and local totals:
Federal ------------------------------------------------------ .2 -------------- .1
State -------------------------------------------------------- 24.6 1.7 22.8
Local -------------------------------------------------------- 28.7 8.8 26.7
Police and fire ------------------------------------------------ 6.9 24.1 6.7
Teachers ---------------------------------------------------- 28.5 26.9 30.0
Teachers (higher education) ----------------------------------- 3.4 3.9 3.9
Other ------------------------------------------------------- 7.6 34.5 .9.7
Total ------------------------------------------------------ 100.0 100.0 100.0
Note: Data relates to table 14 in app. 1.
SINGLE AND MULTIPLE EMPLOYER PLANS
As might be expected nearly all of the smaller state and local public pension plans are maintained by a single employing unit of government (see Appendix 1, Table 5). However, over 40 percent of both the federal plans and the largest state and local plans are maintained by multiple governmental units. Additional characteristics exhibited by "multiple employer" plans are discussed in the next chapter.
SOCIAL SECURITY COVERAGE
In 1975 45 percent of the employees in plans maintained by the federal government and its agencies and instrumentalities were covered under Social Security. The Military Retirement System with 2.1 million employees is the largest public system providing benefits supplemental to Social Security. Among the other federal retirement systems serving to supplement Social Security are the Federal Reserve Bank Plan, the Tennessee Valley Authority Plan, the Federal Home Loan Bank Plans, the Farm Credit District Plans, and the Nonappropriated Fund Plan (see Table B2). The 2.7 million employees contributing to the Civil Service Retirement System make up the largest single group of employees presently excluded from Social Security. Also excluded are the President, Vice President, Senators, Representatives, and federal judges. In total about 55 percent of the employees in federal plans are excluded from Social Security.
Compared with the federal participation, a larger percentage (70.1 percent) of state and local employees in staff retirement systems are covered under Social Security. Table B7 shows that state and local employees working full time are twice as likely to be covered under Social Security as are part-time workers. While nearly 85 percent of all state employees have Social Security coverage, only 36.4 percent of the niembers in police and fire plans bave such coverage. Teachers are less likely to be covered under Social Security, 56.5 percent, than are other local government employees, 75.9 percent. University faculty are as' likely to have Social Security coverage, 83.8 percent, as are other state employees.






'59
TABLE B7.-PERCENTAGE OF ACTIVE EMPLOYEES COVERED 13Y SOCIAL'SECURITYI
Percent of employees
Less than
Full-time full-time
Employee category employees employees Total
Federal plans ------------------------------------------------ 46. 8 8 45. 4
State and local plans:
State------------------------------------------------- 84.9 118. 1 84. 9
LocalI--------------------------------------------------- 76.0 100.0 75.9
Police and fire -------------------------------------------- 37.2 22.2 36.4
Teachers ------------------------------------------------ 62.6 40.6 56.5
Teachers (higher education)------------------------- --------- 85.9 30. 5 83. 8
Other -- ------------------------------------------------ 78.2 10.2 86.6
Total.------------------------------------------------- 72.2 31.5 70.1
Note: Data relates to table 14 in app. I.
There are many important issues arising from the coverage or noncoverage of public employees under Social Security. These issues include universal Social Security coverage, the "integration" of public pensions with Social Security, and pension portability (see Appendix XIII for a comprehensive bibliography on these topics).

PENSION COVERAGE NOT UNIVERSAL
The information in the preceding, section is based on those public employee retirement systems included in the Pension Task Force Survey. 'Not all public employees, however, are covered under a pension plan. The Bureau of the Census reported that in October 1075 about 12.1 million persons were employed by state and local governments on a full or part-time basis. Since 10.3 million employees were reported as having pension coverage in the PTF survey, between 1 and 2 million public employees remain without pension coverage. This finding is consistent with that from the 1972 Ceinsus of Governments in which approximately 11 percent of all full-time public employees were shown to have no pension plan coverage (see Table B8). However, about 78 percent of those full-time employees without pension coverage were shown to have Social Security coverage. A higher percentage of part-timae employees, probably more than 50 percent, do not have pension coverage. (See Chapter D for more information on pension plan eligibility provisions excluding part-time employees.)













74-365--78-5











60



c" em r % 1 9"" co
c wwwww MCI, W Ow cn (n Go cm co
is cp E
C L



CL
Cm fl- d, 00 40 M 0
"f.O=Ln w C" ull) Z6
P, M ull) CD
vl= LO
o ow
CD
c E
un, cc
to cn m"mfcm Ar r'. 00 V* v..
-4 Ln M M r to
-% -cr ull


tm
0 C.J r%.
v-4 4= m U")
S
ui CL

C/)
CD
P Ln M 00 P- 00 wr q**
-4 Uj --4 cn 04
C) 00 00 cn Zen "M

cr

w 00
cr- r W Ln %a W W Ln
gn 0 1*- LO
ui
w
>
0 C fl04 t6 %_O U") r- Lm 4= r,- U*j M d. M,..q 00 cz
-t:t en 00 00
E 0 1" -:21
.0 f.0 cn m Clq C> cc m
U') Cj cm 0 'o W -,

;i *,
CD 04
M W LO -,r U)
to 4-,v
Ln CO C-4 lo. 00
cm
1 cli
CL ca c'!
40
C
00
co 0
ul) cn -e P, r-- C-) P. 00 M P. -4

CL

C-) co
uj a) w 0
to W CD -tr -4 uIll m
U") 00 ca m C-4 CC U) m 6--.W 'm
C, an
CIJ U- -4 -.4 00
<
wt >
4W 0
v = "j
r.% LO
47)
LO
cc




mr w
ui 0 C) Cn Ln -4 -4 -4 cn cn Ln M 00 4)
"CC

0


4W.:c
00 rl
-W 00
Ln to >
Crt cc
uj I:li to=
> E


Ln r-. r C) LO W
> 0) U") ul)
00 Oo >
CY) Q>
p
C,
(D CD CD 00 ",Rt
00 ( 00 a,
E cn f--"M-w U-) (D -47* CJD Ul Cj r*r- CIJ 00 "r Clq C) CJ ce) 00 rci'ci 6 CNT C4 Ocr
QO)
LOCOMWC, a; W 06
to C"i 00 to 00 R**
0.00 fn C Cj to 14 :e ,,,
LLJ (Dt -ts, CO) :t U*) C:) V,
0 *--l' -.400-4 -4 CD
C4 CT
E CD to ;", E

LL. IE -0
0
uj
00 CD
C-7 0 (n 00 00 M
< cn cc
CD f., 00 w o
ui > 00 00 Wm a
0 E > E 8-a






fA W.V7


> c4
0 CIO
Q3 Qn

>,cm

> 41 0 O"a >,,
0 1 Ict
boa CD
> 0
0 to
000 o(>Q. 0 W

>
w
1 .5












CHAPTER C--PERs ADMINISTRATION, REPORTING, AND DISCLOSURE
The Pension Task Force survey documents the wide variation in administrative and operational practices found in the public employee retirement system (PERS). This study is not the first to find public pension plan administration and the reporting and disclosure of plan operations to be nonuniform and, in many cases, inadequate.' One expert on municipal finance has observed that "high on the list of reason for long inattention to pension matters is the paucity of useful information about the status of most jurisdictions' pension systems"1.2
This Chapter discusses the development of the public employee retirement system and the fact that the PERS has generally evolved in the absence of an overall policy framework. It also discusses retirement system administration, the wide variations in pension plan control, the infrequency of routine audits of public plans, and shortcomings in disclosing important plan information to participants and beneficiaries. Lastly, this Chapter discusses the general state of confusion surrounding the qualification of public pension plans under the Internal Revenue Code, as well as certain deficiencies noted in
recordkeeping and other procedures.

PERS DEVELOPMENT AND CONSOLIDATION
The development of the public employee retirement system began earlier than the private system. In 1857 New York State passed a law providing lump sum benefits to New York City policemen injured in the line of duty.*It took another 21 :years before the city's policemen were able to retire without proof of incapacitation-half final pay at age 55 with 21 years of service.' Even before the 1900's, groups of policemen, firemen, and teachers were. covered under service-related retirement systems in New York, Boston, and other cities. Over 12 percent of the largest State and local plans in current operation were established before 1930 (see Appendix I, Table 2). The federal system follows a similar pattern with service-related pensions for military personnel having first been paid in,1861 followed by the establishment of the Civil Service Retirement System in 1920 for civilian personnel.'
The number of large public plans increased dramatically during the period just prior to the enactment of Social Security legislation in 1935 but before optional coverage was afforded state and local government employees. Nearly one-half of the largest state and local plI ans were established during the period 1931 to 1950 when Social Security
1For example, see California's Public Pension System: An, Interim Report to the California Legislature by the Senate Subcommittee on the Investment of Public Funds (Senator Alan Robbins, Chairman), November 1974; and Florida's Local Retirement Systems: A Survey, published annually by the Florida Department of Administration, Division of Retirement.
2 Evalu~ating the Fiscal Significance of Public Pension Liabilities: Somne Suggestions for Municipal Credit Analysts, by Dr. Bernard Jump, Jr., prepared for presentaticn at the Public Securities Association Public Finance Conference, Marco Island, Florida, October 6, 1977.
3 Tilove, Robert. Public Employee Pension Funds. A# Twentieth Century Fund Report. Columbia University Press, New York and London, 1976. pp. 261-262.
(01)





62

coverage for public employees was being debated. Over one-third of the largest plans were started or restructured in a major way after 1950 when state and local employees were extended the option to join Social Security. In contrast nearly two-thirds of the small plans were started after 1950; nearly one-fourth since 1970. (Data showing the dynamics and continuing development of the PERS axe presented in Tables 3 and 4 of Appendix I.)
Over 40 percent of the larger state and local plans have increased their scope of operations in the past by adding new employee groups to their coverage. The period since 1970 has been most active one with over one-fifth of all plan "absorptions" having taken place in this period. Considering this trend to consolidate smaller plans into larger ones, it follows that the public plan universe may have been significantly larger only a decade ago.
When old plans axe restructured or consolidated, membership is usually closed and new hires covered under the new plan. Such "closed" plans now make up about 6 percent of the public plan universe.
Sometimes the reverse of plan consolidation occurs among public pension plans. Nationwide 3.4 percent of the largest local plans now in operation were at one time part of an even larger local or statewide system. In some state-administered plans, such as the Pennsylvania Municipal Employees Retirement System, local employee units have the option to withdraw from the system and on occasion have done so.
Many new plans have consolidated operations in an attempt to rationalize benefit policy and to gain expertise and efficiencies in plan management, accounting, actuarial, computer, and investmeDt functions. Plan consolidation has resulted in administrative savings and proved beneficial in some cases, but there can be barriers as well as some disadvantages to consolidation.4 The creation of one large plan covering many different groups of employers has by no means proved to be a panacea for all public pension plan problems.6
The issues surrounding the consolidation of plans for two or more different groups of public employees are complex. Consolidation usually results in what might be called a "multiple employer" pension plan which can take many different forms. The Civil Service Retirement System is a multiple employer plan covering the various Federal agencies with a unique aspect. being that a local government employer-the District of Columbia-contributes to the plan on behalf of its general government employees.
For purposes of the Pension Task Force survey, a plan being administered in most respects on a separate basis was not considered a multiple employer plan merely because plan assets were commingled with the assets of other plans for investment purposes (e.g., the Illinois General Assembly, Judges, and State Employees' Retirement System pension funds are managed by the Illinois State board of Investment, yet each fund was considered a separate "plan" in the survey). On the other hand, a centrally-administered system was
4 Many of the advantages and disadvantages of consolidating are given in: Retirement System Consolidation: The South Dakota Experience, by James Jarrett and James E. Hicks, The Council of State Governments, Lexington, Kentucky, November 1976.
6 For example, see the Commonwealth of Pennsylvania State Employees Retirement Syst8m Auditor General Report of Examination for the period June, 1973 to December 31, 1974, (including press release); see also an article on the same system in the Wednesday, June 6, 1973 issue of the Irwin Standard Observer, p. 8.






063

considered'to be a multiple employer plan even though all financial, actuarial, contribution, and other plan aspects were accounted for separately for each contributing employer (e.g., the Missouri Local Employees Retirement System). See Table C1 for the distribution of single e and multiple employer pension plans and the accounting treatment of plan finances under multiple employer plans.

TABLE Cl.-ACCOUNTING TREATMENT OF SYSTEM ASSETS, LIABILITIES, RECEIPTS, AND DISBURSEMENTS FOR:
SINGLE AND MULTIPLE EMPLOYER RETIREMENT SYSTEMS 11n percent of plans]

More than 1 contributing employer
Separate
accounting
for each
employer;
assets
allocated
so as not
to pay
Accounting Separate benefits to Multiple
Single on a accounting employees employer
contributing planwide for each of other plans System category employer basis only employer employers subtotal Total

1. Federal Government ------ 58.2 21.8 9.1 10.9 41.8& 100
State and local, government:
1I. By size of system:
A. Large------------- 58.7 26.1 6.0 9.2 41.3 100
B. Medium ------91.9 2.7 4.2 1.2 8.1 100
C. S mal------------- 97.2 2.1---------------- .7 2.8 100
Total ------93.5 4.0 1.1 1.4 6.5 100

Note: Data relates to tables 5 and 6 in app. 1.

PUBLIC PENSION POLICY

With notable exception, most jurisdictions, at all government levels, have not developed an overall policy framework to serve as a guide for dealing with public pension issues. There are many reasons for this shortcoming. Historically, the PERS has taken shape through the addition of a patchwork of laws and programs, creating complexity and confusion. In addition, the specialized nature of pension operations often requires those setting policy to have a -technical knowledge of accounting, actuarial, and investment concepts. There is little doubt as to the reasons why public officials and the public at large have in the past been less than' enthusiastic in wading into the public, pension morass.
A unique aspect of the public sector in dealing with pension matters, is the reluctance or inability of one legislature or administration to commit its successors to a particular policy or course of action (beyond whatever "'Policy" is inherent in applicable laws). The lack of a public pension policy and its ultimate effects on .pension program evolvement was a topic di 'scussed ina speech of state Representative Dan Angel of Michigan delivered before the National Seminar on Pension Issues in Washington, D.C. on September 15, 1977.
Right now we have what amounts to a porkbarrel and piecemeal approach to pension modification. We modify one system without regard for fiscal consequences and then other systems want the same. This takes place in a totally poli-






64

tical atmosphere without any regard for how the bill will be paid, by whom, and when. There is a total absence of logical structure. Employees had better get concerned that there is enough cash on hand to meet retirement needs and taxpayers had better get concerned with these massive and increasing debt obligations. We simply cannot continue in this helter-skelter fashion.
Some states have formed retirement commissions to enable legislatures and administrations to deal with pension matters more effectively on a long-range basis. Ten states have formed permanent retirement commissions while other states and the federal government have from time-to-time created temporary retirement commissions.6
Retirement Commissions have had varying degrees of success in dealing with the pension problems giving rise to their establishment.7 Just identifying the problems can be a major task. The Pension Task Force found that states with retirement commissions were more likely to have catalogued information on the plans operating within their jurisdictions than other states, many of which were found to have little knowledge of the number and types of plans operating
within their boundaries.
Generally, retirement commissions set goals 8 which are intended to meet public pension problems such as those documented by the Pension Task Force survey-e.g., the lack of meaningful financial and benefit disclosure; the lack of auditing and actuarial standards and reports; inequitable benefit structures; rapidly escalating pension costs resulting from inadequate past levels of funding; and the complexity and confusion caused by divided pension responsibilities among different jurisdictions and between different agencies or branches of the same jurisdiction.
This problem does not exist just at the state and local level. In an August 3, 1977 General Accounting Office Report to the Congress, the Comptroller General had the following to say about the federal retirement systems:
The Congress has not provided an overall policy to guide the development of federal retirement systems and should do so. Centralization of committee jurisdiction over all federal employee retirement systems would facilitate the establishment and implementation of such a policy. The systems have developed on an independent, piecemeal basis, causing inequities and inconsistencies, as well as common problems. Many of the differences are without apparent explanation.

6 States with permanent retirement commissions as of December 1976 are Illinois, Louisiana, Massachusetts, Minnesota, New York, Ohio, South Carolina, South Dakota, Tennessee, and Wisconsin. More information is contained in Permanent Legislative Retirement Commissions, published by the Special Subcommittee to Study Michigan's Retirement System of the Michigan House of Representatives Committee on Retirement, December 1976. At the federal level, President Carter created the President's Commission on Military Compensation on June 27, 1977 and announced in June 1977 that he was going to create a Presidential Retirement Commission to study private and public pension issues.
r For example, see the recommendations made in the 1973 Report of the Illinois Public Employees Pension Laws Commission.
8 For example, the following pension reform principles were adopted in 1977 by the National Conference pf Stale Legislatures' Task Force on Public Pensions:
1. Creation of permanent legislative commissions with staff and actuarial assistance with responsibility for recommending legislative changes.
2. Requiring all public funds to report annually to the legislatures on a uniform basis.
3. Prohibiting changes in pension benefits or contributions by any body other than the state
legislature.
4. Statutorily prohibiting collective bargaining on public pensions.
,. lImposing a moratorium on any reduction in the age of retirement.
Requiring competent fiscal notes on all proposed pension legislation.
7.Eliminating pension-hopping and double-dipping.
8. Consolidating state and local government pension systems into one plan.
9. Periodic re-examination of disability roles.
10. Requiring legislation which would increase pension benefits to also contain front-end funding on
a sound actuarial basis.
11. Integrating all state and local systems with social security (since the taxpayer is the basic employer
of any government pension system).





65

RETIREMENT SYSTEM ADMINISTRATION
While the survey data show that larger plans are generally better managed than smaller plans, in many c ases largrerDlans display the same administrative weaknesses as smaller plans. In addition, large multiple employer plans have been found to have some unique problems of their own. The survey data supplements the findings of a state ,official who discovered municipal pension systems in one state to be "in a primi'tive and discordant state". 9
With rare exception, the administrative responsibility for large public employee retirement systems is vested in a retirement board (board of trustees, investment board, etc.). Generally, retirement boards are established under provision of law and axe, therefore, responsible to legislative bodies or elected officials. In contrast, plan administration in nearly one-third of all smaller state and local pension plans is carried out under the offices of elected public officials. These officials may in some cases turn over administrative duties to persons outside government.
The extent to which control over pension policy is delegated by statute to a retirement board varies from system to system. In some large systems, retirement boards make legislative recommendations, set investment policy, establish rules of benefit entitlement, review or approve administrative budgets, and hire inside staff and outside consultants to carry out day-to-day operations. At the- other extreme, retirement boards may be established only to review. disability and retirement claims. In some cases legislative bodies may retain close control over administrative budgets through the appropriation process but delegate investment and other administrative functions to governmental departments, or offices different from the retirement board.
Large city and state systems with broad administrative powers usually appoint a full-time executive director (plan administrator) to, guide the board and to carry out such duties as the board may delegate. As shown in Table C2, 36 percent of the smallest plans and 21.2 percent of the medium sized plans do not have retirement or investment boards. In these smaller plans, city clerks, budget officers, even police chiefs may spend part of their Urne on pension plan adramilstration and investment matters. Consultants, insurance agents, or insurance companies perform some or all of the administrative services for a majority of the smaller plans not having retirement boards.
Hearing before the Subcommittee on Labor Standards of the Committee on Education and Labor, 94th Congress, ist session, on H.R. 9156, Washington, D.C., September 17,1975, testimony of William 1K. Mleox, Secretary, Pennsylvania Department of Community Affairs, p. 70.





66
TABLE C2.-RETIREMENT SYSTEM ADMINISTRATtON
Ultimate policy and administration authority
vested in (percent of plans)
plans havingNeither a
retirement or
Retirement Investment investmentSystem category board board board
1. Federal Government -------------------------------------------- 43.7 9.1 54.6
State and local government:
11. By size of system:
A. Large ---------------------------------------------- 88.5 10.1 11.01
B. Medium -------------------------------------------- 78.2 5.6 21.2
C. Small ---------------------------------------------- 64.0 .7 36. G
Total -------------------------------------------- 67.9 2.1 32.0
Note: Data relates to table 7 in app. 1.
One of the duties for about one-fifth of all retirement boards is to serve as the custodian of plan assets. Individual board members -in 5.6 percent of all plans serve in this capacity. The role of custodian is given to the plan administrator in 3.3 percent of all plans- In 23.5 percent of all plans insurance companies receive plan contributions, and invest plan assets. For the largest number of plans, 42 percent of the total, the treasurer of the related governmental body serves as custodian. Banks and trust companies are assigned custodial duties in 21.4 percent of all plans (see Table 9 of Appendix 1).
Regardless of their composition, most retirement boards exercise full authority in investing plan assets (subject to applicable law as. discussed elsewhere in this report). Increasingly, retirement boards. have been turning to both in-house and outside professional advice and management in the investment area. As shown in Table C2,. over 10 percent of the larger plans, mainly at the state level, now turn all investment functions over to unified investment boards. Among other states, Illinois, Minnesota, and Wisconsin have established separate investments boards at the state level.
The effectiveness and sensitivity with which a retirement board can deal with complex administrative and investment issues is dependent on the interest, experience, and abilities of its individual members. Detailed insight into the characteristics and composition of retirement boards can be obtained by referring to Table 8 of Appendix 1. This table shows that employee representatives constitute a board majority in about one third of all federal, state, and, local systems. On the other hand, employees have no board representation in 28 percent of the plans. The retirement boards for small plans and local plans are more likely to have employee majorities than are large state plans. This is partly accounted for by the fact that police and fire plans having em Floyee board majorities in nearly 40 percent of the cases make up a a geDercentage of small local plans.
Employee board repre'sen'tatives are more than three times as likely to obtain their board position through an election process than by appointment. Employee representatives are elected by other -plan pai-ticipants in about 83 percent of the local plans but in only half' that percentage of state plans.






67

In another one-third of all systems, elected and non-elected government officials control plan administration through collective board majorities. There is a sharp -contrast between state 'and local plans in this regard. Governmental officials have board majorities in about 37 percent of the local plans while obtaining majorities in only 14 percent of the state plans. In the vast majority of cases, government officials achieve board membership on an ex officio basis. Only in a small percentage of plans are they elected by plan members. In about one-fifth of all plans, retirement boards have been established without any representation on the part of elected 'or other government officials.
An important difference between state and local plans is the large percentage-nearly '23 percent-of state plans having retirement board majorities consisting of persons employed outside of government in fields unrelated to investment or finance. The comparable percentage for local plans is 2.7 percent. Board membership for such persons is usually attained by appointment. On the other hand,' less than I percent of all retirement system boards have members who are persons employed outside of government in fields related to the investment, banking, or finance field.
As diligent as some administrators have been in fulfilling their duties, the lack of clear-cut pension policies, direction, and control (discussed in 'earlier sections) has prevented 'some systems from overcoming certain administrative weaknesses. This has been most apparent in the administration of the disability programs of some systems, large and small.
Systems having similar service and disability retirement provisions covering employees of the same occupation show wide variations in the relative number of disabilities granted.
On a national basis the number of disability retirees in the police and fire category as a percentage of all retirees is about 23 percent. Yet for large similarly situated police and fire plans this ratio ranges from less than 10'percent to over 80 percent. A large share of such variation which cannot be attributed to different definitions, environment, etc. results from varying degrees of what might be described as administrative largess. Administrative laxity in the disability area has forced at least one plan in the past into court appointed receivership.'0 Small plans can be particularly yulnerable to abuse inthis area.
State-run plans generally exhibit lower ratios of disableds to total retireds than do city-run plans. The overall ratio for plans in the state and local government category is 8.5 percent as compared with a ratio of 26 percent for the federal Civil Service Retirement System. While the definition of disability for the federal system is 'more liberal than that for many state and local plans, a part of the 213 percent difference is undoubtedly due to differing administrative procedures.'"
The seriousness of the equity and cost implications of inadequate administration in the disability area has begun to be recognized in at least some federal, state, and local systems.'12
1See New York Times, March 1, 1972, "Hludson County New Jersey Plan Put Tnto Receivership."
1Repo'rt to the Con gress by the Comptroller General of the United States, Civil Service Disability Retirements: iVeeded Improcemen's, November 19, 1976.
12 For example, see Report No. 94-1728, 94th Congress, 2nd Session, September 29, 1976, on the District of Columbia Retirement Reform Act, p. 9-11.







In other benefit areas, administrative procedures have proved inadequate to prevent certain benefit provisions from resulting in favoritism, other inequities, and abuse. Over one-fourth of all public plans calculate pension benefits on base pay plus overtime pay during the last year (or last day) before retirement. The manipulative pay and personnel practices used to achieve larger benefits under plans with such provisions are well-known.13 Less attention has been paid to another provision of some public plans which gives part-time employees the same service credit as full-time employees. A part-time employee can then receive the same pension as a full-time employee with the same service by being allowed to work full-time for the final 3 or 5 years before retirement (the period used for computing benefits). For this and other reasons many public plans were unable to distinguish between part-time and full-time members for survey purposes (see footnote to Table 14 of Appendix I).
The task of the public pension plan "administrator" is a difficult one given the number of different parties exercising direction, control, and influence over public pension matters. The time and attention of the plan administrator is demanded by legislative bodies, elected officials, various boards and commissions, employee representatives, and other special interests. Even where one person is given the title of plan administrator, as is the case for most large plans, plan administrative functions (including investment ones) may be carried out by persons working in private or public capacities under various offices, departments, agencies, boards, or institutions.
The answers administrative personnel give to the question "Who has ultimate plan control?" are instructive and lend insight into the special nature of governmental plans. The ultimate authority in different cases may reside in the mayor, the city council, the board of education, the police or fire commission, the governor, the state legislature, retirement or investment boards, the U.S. Congress, or even collective bargaining agreements. For the most part, but not in all cases, federal, state, and local statutes or codes govern the establishment and operation of public plans. To the extent such laws do not exist or are less specific in their requirements, the form and structure of plan administration is left to the discretion of those responsible for administering the plan.
The controls over'pension plan administration, whether by law or otherwise, are complex and confusing and have proved inadequate in some cases. Some very small pension plans have a retirement board with more members than the number of participants in the plan.14 A lack of adequate control leads to abuses, such as occured in one community having four police chiefs in six days in order to entitle the officers to higher pension benefits.'5 Pension plans administered without the benefit of statute or even written plan documents are an open invitation to abuse.'6
13 For example, see Hearing before the Subcommittee on Labor Standards of the Committee on Education and Labor, 94th Congress, 1st Session, on H.R. 9155, p. ,0; also see The Evening News, Harrisburg, Pennsylvania, October 13, 1976, "40 CI ies Face Agony on Pension-fund Handling.' Also see New York Times, July 27, 1975, "New York City Transit Workers Increased Benefits Because of Overtime." 14 Report of the Illinois Public ,mployees Pension Laws Commission, 1973, page 17. 15 Ifearing before the Subcommittee on Labor Standards of the Committee on Education ad Labor, 94th Congress, 1st Session, on H.R. 9155, Washington, D.C., September 17, 1975, testimony of William 11. Wilcox, Secretary, Pennsylvania Department of Community Affairs, p. 70. ] Ibid., p. 73.







While the absence of specific legal requirements pertaining to
pension administration may leave open certain, avenues leading to manipulation and abuse, a maze of laws built up over time may add another form of conflict or confusion. For example, in one state "the provisions . have become so vague and confusing that administrators of the various systems must select one of a variety of interpretations and hope they are right." 11 In other states, the state legislature may increase benefits for local employees while leaving to the local jurisdictions the job of raising revenues to finance the new benefits. A chaotic condition can result when local jurisdictions cannot raise the necessary revenues because of state. laws placing a limit on local rates of taxation.'8 Local jurisdictions can also be placed in a state of financial uncertainty when state laws stipulate contribution rates, which prove inadequate.'9
Recognizing the confusion and conflict that a patchwork of laws can create, several states, have taken steps to provide greater uniformity and control over pension matters.20 However, increased State control has not necessarily solved all the financial and administrative difficulties found in those states. Even where some States have experienced greater control, rising pension outlays and swelling pension liabilities continue to reflect inadequate pension funding in many States .21 In addition, rarely do the states retain control over practices which have great. effect on pension costs such as salary (for final average pay plans), personnel practices (e.g., granting overtime), and collective bargaining (now 'extending to about 35 percent of all em-ployees, covered under 22 percent of all plans) ,22

RETIREMENT SYSTEM AUDIT$
The most striking characteristic of, public employee retirement system audits is the absence of any uniform or standard auditing practice. AboutA4.6 percent of all state and local plans and 29.1 percent of all federal plans are not audited at all (see Table C3).
A substantial segment of the public employee retirement system is characterized by the. absence of any regular or external independent review of plan operations. Nearly one-third of all state and local plans and 37 'percent of the larger ones do not provide for annual audits. Only 47.4 percent of all plans and 38.9 percent of the larger ones are audited annually by licensed or certified public accountants outside of government. Another 10.8 percent of all plans are audited by independent accountants, but not on, an annual basis.
17 Recommendations For Action on Local Retirement Plans, State of Michigan Department of Management and Bud 'get, Office of Intergovernmental Relations, 'May 1977, p. 7; another example of differing interpretations of state law regarding benefits is in Pennsylvania Department of Community Affairs Reports, May 1976, article by Conrad M. Siegel, p. 2.
18 Hearing before the Subcomamittee on Labor Standards of the Committee on Education and Labor, 94th Congress, 1st Session, on H.R. 9155, Washington, D.C., September 17, 1975, testimony of William H. Wilcox, Secretary, Pennsylvania Department of Community Affairs, p. 64. 19 Denver Post, December 8, 1976, "Englewood Sues State Over 'Unsound' Pension Law." ;0 Among others, most of the states with permanent retirement commissions (listed in footnote 6) have taken steps at the state level to achieve greater uniformity and control over state and local pension matters. 21 Illinois and Massachusetts are only two states where retirement commissions have called attention to the need for increased pension contributions to budget and finance pension costs on a current basis. 22 For example, a city participating in one statewide system attempted, in this case unsuccessfully, to permit employees to convert insurance and other city paid benefits to "income" for the purpose of boosting pensions by as much as 20 percent. See Pensions & Investments, February 28, 1977, "Madison Loses Rebate Money for Trying to Boost Employees'Icomes," p. 10.







70

TABLE C3 -EXTENT OF RETIREMENT SYSTEM AUDITS

Percent of plans
Audited, but not
Audited every year annually
Not Outside Outside
System category audited audit Total audit Total Other Total

1. Federal Government---------------- 29.1 40.0 50.9 1.8 5.5 14.5 100
State and local government:
If. By size of system:
A. La rge------------------ 1.3 38.9 63.0 11.3 32.4 3.3 100
B. Medium ---------------- 7. 3 34.6 51.4 16.7 30. 7 10.6 '100
G~. Small------------------ 4.4 50.7 70.6 9.6 18.4 6.6 100
Total ---------------- 4.6 47.4 67.0 10.8 21.4 7.0 100

Note: Data relates to table 11 in app. I.

Characteristic responses for the plans shown as "other" in Table C3 were: "audited once in 13 years by CPA firm"~, "last audit by
state was in 1971", and "funded by insurance". The remainder of the plans not previously discussed are audited periodically by agencies of federal, state, or local governments. About 45 percent of the largest plans and 30 percent of all plans fall into this category. Some, of the largest public pension plans in the country are only audited every 4 or 5 years, and they are audited at that time by related governmental agencies.
Some pension experts have questioned the appropriateness and adequacy of public pension plan audits by related governmental agencies. The executive secretary of one large state system had the following to say about audits by state examiners:
In addition to state politics, other serious limitations of State Examiners exist. First, State Examiners generally lack experience and expertise in the specialized needs of the public pension fund organization. Principally, their auditing expertise lies in the expense and departmental operational accounts that revert to the state, and not organizations that carry over hundreds of millions of dollars each year. Second, State Examiners are seriously limited in manpower and budget, thus allowing only 2 or 3 auditors on a particular audit. It should be quite obvious that a complete audit demands experts in such fields as administration, investments, data processing, actuarial sciences, records management, as well as the traditional accou nting techniques and methodologies.
Financial institutions, such as banks and insurance companies are audited by as many as 4 or 5 different private and/or public agencies, and they are continuously finding shortcomings in management and performance. In my opinion, there cannot be enough audits when it comes to keeping a close eye on hundreds of millions of dollars being invested for the vital needs of hundreds of thousands of citizens of your respective states.23
The lack of an independent review of public pension plan financial an(] actuarial matters carries an attendant risk of financial miscalculation or abuse. In regard to the need for independent review, public pension pl.3ans cannot be viewed differently from other financial enterprises including the sponsoring govern mental employers thems elves. While related to the area of municipal finance generally, the following statement in regard to New York City's recent financial problems made at a news conference by Comptroller Harrison J. Golhlin could have equal significance in the public pension"'plan con text:

2 Article by Dr. David G. lBrornier, Rdirement Sy.,tem8 of Alabama Advi8or, Vol. III, No. 12, July 1977#








Most important of all, what have we learned from.the ordeal of this period?
1. We have learned the risk when there is no outside, independent review and oversight of the city's condition, the kind'of independent review that began with the audits I released in the summer of 1974, We must be sure that such review and oversight continue to exist.
2. We have learned that the shambles and chaos of the city's accounting system, which I revealed, made it impossible to maintain effective control or to secure reliable data. Uniform accounting standards must be required for all municipalities -as a matter of law.14
A recent Coopers & Lybrand-University of Michigan report on Financial Disclosure of the American Cities concluded that compliance with voluntary standards relating to accounting, auditing, and financial reporting practices is ineffective
Our findings indicate a substantial lack of compliance with current generally accepted principles applicable to governmental bodies. These findings support our recommendation that disclosure compliance for the protection of taxpayers and security investors should be accomplished through uniform enforcement. The MFOA [Municipal Finance 0 ffilcers Associatio.-Li] recommends compliance with GAAFR [Governmental Accounting, Auditing and Financial Reporting]. However, the organization issues fewer than 40 Certificates of Conformance each year. In fact, only about 400 such certificates have been issued over the last 30 years. Out of approximately 18,000 municipalities eligible to apply, only about 100 applications are even submitted annually for consideration. This points up the hopelessness of voluntary compliance.25

PLAN DISCLOSURE TO PARTICIPANTS

The persons administering public employee retirement systems are rarely obligated to carry out their fiduciary dutie s "solely in the interests of plan participants and beneficiaries." The loyalties of plan administrators at the federal, state, and local levels are divided, atvarious times and to various degrees, between authorizing or appropriating legislative bodies, elected or appointed executive officials, various boards, and the special interests of board members including in most cases elected or appointed representatives of plan participants. Plan administrators may also be confronted with conflicting and confusinz statutes and court interpretations creating legal uncertainties as to plan provisions and operations. The above factors as well as others have contributed to a wide-range of public pension plan disclosure practices.
In many cases plan disclosure to participants is inadequate or nonexistent. In such cases plan participants and beneficiaries (and other interested parties as well) are unable to assess properly plan financial operations, are unappreciative of the true level of pension costs, and are unaware of conditions leading to benefit losses. There is an increasing recognition on the part of public pension official-s of the need for additional disclosure of public pension plan financial and funding
26
operations. Later chapters deal with present plan practices in these particular areas.
In various ways some of the largest public employee retirement systems give recognition to the impoAance of meaningful disclosure in furthering participant understanding of oftentimes complex benefit

24New York Times, August 29,1977, p. 20.
25 Financial Disclosure Practices of the American Cities, Coopers and Lybrand and the University of Michigan, Washington, D.C., 1976, p. 37.
26 For example, see the recommendations of Mr. Robert E. Blixt, of the Minnesota State Board of Investment, in Pension World, Vol. 12, No. 12, December 1976, p. .-4.






1 72

provisions. The plan description of one large system instructs plan participants in the following manner: "You are urged to acquaint yourself with the information contained herein and to review it from time to time so that you w,7ill be familiar with the obligations and rights, and can plan intelligently for the future."
Survey data suggests that participants in a majority of all public pension plans are hindered in their ability to "plan intelligently for the future" clue to inadequate disclosure of plan provisions. Public pension plan disclosure practices may be tested against any of several basic principles of effective communication: (1) that plan materials be read ily available; (2) that the materials be timely and up-to-date; (3) that tihe materials be comprehensive so as not to omit important plan features; and (4) that information be presented in a manner that will be understood by plan participants and beneficiaries.
While nearly 85.percent of all state-run plans automatically provide plan participants with booklets or other materials describing plan provisions, onl y 42.2 percent of locally administered plans so provide (see Table C4). The high percentage of teacher plans, 92.7 percent, automatically furnishing plan descriptions is noteworthy. While all or nearly all university and college faculty and others participating in ,plans funded through Teachers Insurance and Annuity Association of America-College Retirement Equities Fund (TIAA-CREF) automatically receive plan descriptions, about 61 percent of the large teacher plans administered at the state level automatically distribute plan descriptions to all participants. In contrast, plan participants in over one-fif th of all other plans are unable to obtain plan descriptions even upon request.










P"ICI




CD CD CD
-4 -4





Ln en am m co





.0 4 1 >, kn C%4 P Ln
.,- c & cc 0 CD c
= = 0 C4 06 0; g n; 'i 'd
E =) Cr co cn




w Go 00
cn
r Ln co




co Ln CD C=l wt m
_; C4 cy;
V-4 C-4


CL

(D C> CD CD CD
CD CD CD CD CD C)





E c
CC M 3:
Lo E 06


LU 0
4v.= E =
cn- C14
CL 0 =
= 0 V-.4 Cf) CY) en
= cn
La 0 cr
oz

.7 od 'r; C'
_j CD Ln "m U')
(n Wa' CD
06
to
ui 06
Z
en 00 Cj
4 C4 cr; ui c; od
CD
CL CL


ui

CD
o:: C> 40 CD C> CD
-4 -4
uj
cr.


CL w ,p C-4 C4 On






co C*4 Go r 40
CXw C4 rz C4 06 td ui
W P o V" C14 C14 C14
cr



co cm cm r, to
.4 C4 C-ig j 6
m .E


CL Cj co Cl
2C ad C4 cyi rz 4
C14 C*4 C14
CL
0
S

40 Im ow f 6, 1 I 06C> 1 ="O I

cc

E
M M cc le 0 W=o 0
=-=> 5. W.=
2.2 MI.-* 00 -61B cc to M.16. -A 9 00 CD
9 E
cc
cats

co CA


2C





74

Where plan descriptions are furnished automatically or upon request, there exists great variation in the form and usefulness of such materials. Some of the largest plans distribute elaborate and comprehensive plan "booklets". In other cases descriptive "pamphlets" are so brief as to be of little help to participants seeking answers to practical questions. Some plans, upon request, will furnish verbatim copies of the applicable sections of state and local law which may be incomprehensible to all but the most sophisticated plan participants.
To be useful and to avoid misleading plan participants, plan description materials should be kept up-to-date with plan amendments and other changes. For the most part the public pension plans furnishing participants with plan descriptions on an automatic basis indicate that such materials are periodically updated or rewritten. For some plans, however, substantial periods of time may elapse between plan description updating.
The following example illustrates the potential which is created for benefit losses when plan disclosure falls short of some of the "basic principles" of participant communication given above. The case in point, involving the employees of several federal employing offices, also serves to illustrate the unique problems which multiple employer plans have in dealing with, *in some cases, thousands of separate employers. The plan description for the participant group in question was issued nearly eight years ago and is currently out-of-date in several respects. The document is not distributed automatically, is nearly out-of-print, and the availability of the document is unknown to many, if not most, of the employees involved.
Under plans exhibiting circumstances similar to the above, plan participants may be generally unaware of important retirement features and valuable coverage in the event of death or disability. Participants and beneficiaries may also be unappreciative of the voluntary or mandatory aspects of various portions of their benefit plans, thus forfeiting benefits otherwise attainable under "portability" and "buyback" provisions (see Chapter D), various survivor and retirement options, or optional benefit programs. An inadvertent loss of benefits can also occur when participants and beneficiaries are not made familiar with plan vesting provisions or with remarriage, reemploynient, or earnings limitations that apply after benefits commence.
The contributory nature of most public employee retirement systems creates special needs for plan disclosure and employee understanding if benefit losses are to be avoided. In such plans, many employee benefit rights hinge on the disposition of employee contributions.
A glaring weakness of contributory retirement systems presently is the inadequate means of informing participants of their rights to vested benefits or to the return of their own contributions upon termination. About one-half of all 'federal, state, and local plans (lo not automatically furnish employees with statements of their contributions (See table C4). In over 8 percent of the plans, the participants are unable to receive such statements even upon request. In contrast, nearly all teacher plans (inchi(lincr TIAA-CREF plans) supply their members with individual statements of employee contributions.






75

The results are predictable. Many pension plan membership rolls and trust funds are clogged with the names and accumulated contributions of former members with which the plans have lost total contact. The problem is a multi-million dollar one, and the only way some governmental jurisdictions have acknowledged the problem is by trying periodically to expropriate the unclaimed pension funds under state escheat laws.27 The problem of unclaimed benefits has also been a perennial one for the federal Civil Service Retirement System. A 1972 General Accounting Office report estimated :338,000 former federal employees aged 62 or over as having potentially unclaimed benefits totaling $26 million."8
The federal government attempted to deal with this problem by passing Public Lawv 94-183 which established a time limitation in applying for civil service retirement benefits. Under present law, no benefit will be paid unless former employees apply for benefits before their 115th birthday, and after the death of an employee unless the application is received within 30 years after death. The effect of these provisions is to authorize the Civil Service Commission to destroy retirement records when no claim for benefit has been received within the periods specified by law. While this may solve the recordkeepingproblems, it does not solve the problem for the individuals who may need this income in their retirement years. The GAO concluded that returningg the money would serve a purpose as useful as, or more useful than, correcting a serious recordkeeping problem and would provide benefits expected of the Government." Accordingly, GAO recommended that a program be implemented to return unclaimed retirement funds and that federal agencies be advised of the extent of unclaimed benefits and reemphasize the importance of providing proper counseling to people leaving Government employment so they will be aware of their rights to retirement benefits.
Many employers furnishing individual statements of employee contributions do so as part of-their payroll operations. In such cases employees may be unaware of the importance of retaining payroll "stubs" for retirement benefit purposes. In the absence of effectivedisclosure of plan provisions, employees who quit may forfeit vested employee contributions or other benefits regardless of whether or not individual contribution statements are furnished them while working. The result is that federal, state, and local pension plans have accumulated millions 'of dollars in unclaimed benefits for, perhaps, hundreds of thousands of former employees .29 The fact that the federal -system retains the contributions of approximately two million former federal. employees suggests that the problem of unclaimed benefits will continue for this system as well.
.To avoid such losses and to better inform participants -of their rights, some plans provide employees with benefit statements containing information on accrued benefits or accumulated contributions.
27 N-ew Jersey's Contributory Pvblic Emploee Pe-nsion Proorams: Program Analysis of the Public Empioyeeot' Retirement System, Office of Fiscal Afyair othe Ne* Jersey State Legislature, March 1976. p. viii. 28 Report to the Congjress by the Cornptroller General of the United States, Unclaimed Benefits in the Civil Service Retirement Fund, December 20, 1972.
29 See New Jerseyj's Contributory Public Employee Pe'nsion Prkograms: Program Analysis of the 'Pulic Employees' Retirement System, Office of Fiscal-Affairs of the New Jersey State, Legislature, March 1976.. p. v-vi; also see Repor totecngesb h oprle eeaf the United States, Unclaimed Benefits in the Civil Sf rvice Retiremen t Fund, December 20, 1972.

74-365-78-6





76

About one fourth of all plans, including about 60 percent of the state plans and 20 percent of the local plans, automatically furnish participants with such information (see table C4). Participants in 19 percent of all plans cannot obtain such information even upon request. When participants are not provided information as to the value of their accrued benefits at the time of termination, losses other than those resulting from unclaimed benefits may occur. Participants may withdraw their own contributions, and in some cases may even be procedurally encouraged to do so, thus forfeiting deferred retirement benefits having a -m"'uch greater value than the" lump sum return of contributions.
There is a wide variation in the degree of administrative attention given to supplying information to terminated participants and keeping records related to such individuals. Many plans were unable to separately identify terminated participants from active participants. 1- or large plans providing such information, the number of former employees with deferred vested benefits ranged from less than 5 percent of the number of persons currently receiving benefits to over 100 percent.
Public pension plans of the multiple employer type exhibited special problems in supplying information on the number of terininated participants. To the benefit of employees such plans count service with all contributing employers for purposes of computing any one employee's final pension. Most multiple employee plans also accumulate service without regard to the length of any "break-inservice" (see Chapter D and Table 24 of Appendix 1). Because the presence of such provisions makes it difficult to identify "terminated" participants, disclosure of accrued benefits to such persons is oftentimes nonexistent.
At any one time multiple employer plans are also likely to have less than complete information on the benefit status of a given participant. This is usually due to the retention of employee records at the contributing employer level with the consolidation of such records occurring only at termination, retirement, or other more regular intervals (one result is that some plans are unable to report the total amount of accumulated contributions for a given participant or for the plan as a whole.) Under these plans greater reliance is also placed on the ability of employees to remember or, perhaps, provide proof of applicable periods of past employment for benefit purposes. The chance for errors to occur thus leading to inaccurate benefit computations is further enhanced if participants fail to understand the gisys tem" and the need in some cases to keep records of past service and contributions. The recordkeeping problems are compounded when a retirement system fails to utilize a unique identifier (such as the Social Security number) for each participant. In such circumstances benefit losses have an even greater chance of occurring in the event of the death or mental impairment of the participant.
In summary, the disclosure practices of public employee retirement sKstems at the federal, state, and local levels fall considerably short of t e h' h t dards set for private pension plans under ERISA. While
11arsyean ans may provide participants with E-RISA-like plan descriptions and finances statements, only a small percentage of all






77

public pension plans could presently meet all of the ERISA disclosure provisions requiring: (1) summary plan descriptions written in a manner calculated to be understood by the average plan participant, (2) summary descriptions of material plan modifications,
(3) summary annual reports of plan finances, (4) statements of accrued and vested benefits upon request and at termination,
(5) written explanations of claims denials, and (6) access to all plan documents including financial and actuarial reports.

INTERNAL REVENUE CODE QUALIFICATION
The question as to whether or not public employee retirement systems need to obtain plan "qualification" under section 401(a) of the Internal Revenue Code has been subject to misunderstanding and confusion. Public pension plan administrators and other officials have admittedly been reluctant to voluntarily submit their plans to the Internal Revenue Service for review.30 The minimal enforcement and unequal application of section 401(a) by the Internal Revenue Service as it relates to public plans has also contributed to the general state of confusion.3
The uncertainties of public pension plans as to any obligations they may have under the Internal Revenue Code are highlighted in Table C5. Well over three-quarters of all plans have failed to seek IRS qualification, and about 57 percent of those queried professed to be unfamiliar with the application of the Internal Revenue Code to their plans. About 44 percent of the federal plans, 23 percent of the state plans, and 14 percent of the local plans applied for and received favorable plan determination letters in the past. Over 60 'percent of the determinations, however, occurred five or more years ago. This indicates that even those plans seeking initial qualification usually do not Pseek later rulings when their plans are amended.
TABLE C5.-APPLICATION OF INTERNAL REVENUE CODE QUALIFICATION PROCEDURES UNDER SEC. 401(a) TO RETIREMENT SYSTEMS
[In percent of plans
Received favorable
I RS determination letter- Received
unfavor- Applied
Not Did not In last able IRS but no
familiar apply for 5 yrs Prior deter- deterwith qualified (1971- to urination mination System category process status 76) 1971 letter letter Unknown Total
I. Federal Government ---------- 14. 5 30.9 32. 7 10. 9 ---------- 5. 4 5.4 100
State and local government:
il. By level of administration:
A. State administration._ 18.9 55.6 4.4 18.5 1.0 1.5 ---------- 100
B. Local administration. -- 68.0 16. 5 6.0 8. 2 ----------- 1.3 ----------.100
Total ------------- 57.5 19. 0 5.3 8.6 .1 1.2 8.0 100
Note: Data relates to Table 13 in app. 1
30 Commonwealth of Penntlvania Public School Employees' Retirement System Audit Report for yeare ftded June '80, 1973 and 1974, p. 24.
3 Alvin D. Lurie, Assistant IRS Commissioner (Employee Plans/Exempt Organizations), address before
-the National Symposium on Public 'Employee Retirement Systems, Washington, D.C., September 14, 1977, p. 2.






78

The reasons given by some plans as to why IRS determination was not sought are illuminating:
(1) Does not apply to municipal governments.
(2) Technically not a qualified plan, but IRS treats as qualified plan for tax purposes.3
(3) The University is exempt from income tax so there is no tax advantage.
(4) Told verb ~a ly by IRS representative that system is automatically qualified.
(5) Attorney has advised that the system would not be eligible and therefore need not apply.
(6) The benefits of qualified status would not justify the potential problems generated by conflicts between IRS rulings and state, statutes.
(7) Have not had either the legal or administrative staff to fight the long war with IRS to obtain qualification. IRS will not give preliminary advice on proposals, consequently greatly complicating and raising the cost of obtaining qualification.
About 4 percent of the state plans seeking qualification failed to meet the IRO section 401 (a) requirements. Federal and local plans seeking qualification indicated no problems in obtaining approval. Although inconclusive, the small percentage of state plan applications rejected by the IRS may indicate problems which some state plans have had in meeting the nondiscrimination requirements of IRC Section 401 (a). (e.g. where highly compensated elected or other officials receive benefits of proportion ately greater value than other state employees) 31
The state of confusion over the application of IRC section 401 (a) to public pension plans has been resolved at least for the time being., As given in Appendix X, the Internal Revenue Service position clearly states that Internal Revenue Code Section 401 (a) does apply to public employee retirement systems. For plans failing to meet the provisions of the Code, the tax consequences to the plan's participants and beneficaries as well as to the related trust may be severe. One source, estimates potential tax liabilities exceeding $388 million may be incurred by public p~ension trusts, if they fail to meet the section 401 (a) qualification standards.4 Addressed in part 11 of this report-federal, laws presently affecting PERS -are the various legal aspects relating to the qualification of public pension plans.
As the surveys shows, the participants and beneficiaries of most public employee retirement systems do not benefit from the safeguards of the nondiscrimination, the prohibited transaction, and the other plan qualification provisions of the Internal Revenue Code. The applicability of such provisions in the public pension plan context must be carefully reassessed, however, in order to ensure that future, applications of the same or similar standards do not directly or inadvertently hinder the funding progress of some plans. For example, one state administrator claims, that as a result of an IRS recomnmendation to eliminate plan dliscrimnination, the state set up a separate
.1 Stee Appendix XT for a clarification of the tax qualification status of the Civil Service Retirement, S3yStein aid other Fode'.ral systems.
33 Penisiong & IvtntJune 21, 1976, "Inconsistency of IRS Qualification Policy for Public Funds, is Shown," p. 24.
34 Library of Congress, Congressional Research Service, estimated liabilities calculated as of June 30, 1976.






79
unfunded plan for judges, legislators, and elected officials who were previously covered under a funded plan for other state employees.35 However; at the present time there is little additional evidence to indicate that any of the public pension plans that now operate on an unfunded basis (17 percent) do so as a result of a deliberate decision to avoid the application of section 401(a) of the Internal Revenue Code.

RECORDKEEPING AND OTHER PROCEDURES
Early public employee disability and pension programs started out as extensions of city payroll operations. In a few large cities and numerous small ones, pension operations have not progressed much beyond this early stage. Most large city and state systems, however, have made substantial progress, many in the last decade or so, mi bringing modern management and computer techniques to bear on their pension operations.
Nonetheless, large public employee retirement systems of the multiple employer variety experience some unique administrative problems of their own. Obtaining and maintaining recorded data relating to, in some cases, thousands of participating employers and several hundred thousand employees present major problems for multiple employer plans. Recordkeeping problems are encountered by such ,plans even when elaborate rules, regulations, and procedures are spelled out and made applicable to all participating employers. The lack of adequate staff follow-through or auditing, which may in turn result from a shortage of administrative funds, can result in the loss of plan income or employee benefits. Some plans experience delinquencies in the payment of contributions, and at least one plan has had to deal with the possibility of employer default.36 Problems in obtaining accurate and complete employee payroll and other data serve to undermine the reliability of.,actuarial valuations and of computed contribution rates.37
Multiple employer plans often lack up-to-date information on the membership and employment status of individual employees. Some plans, such as the Federal Civil Service Retirement System, rely on participating employers to maintain employer service and contribution records and may never combine all individual records until benefits are claimed. As a result, such plans may not know the number of terminated vested employees and may have trouble estimating the related effect on benefit costs.

SUMMARY AND CONCLUSIONS
1. PERS development and consolidation
The development of the public employee retirement system began earlier than the private pension system. Over 12 percent of the largest state and local plans in current operation were established before 1930.
35 Pensions & Investments, June 21, 1976, "Inconsistency of IRS Qualification Policy for Public Funds
-is Shown," p. 24.
36 Hearing before the Subcommittee on Labor Standards of the Committee on Education and Labor, 94th Congress, 1st Session, on I.R. 9155, Washington, D.C., September 17, 1975. p. 31C-317.
3 For example, see the Commonwealth of Pennsylvania State Employees Retirement System Auditor General Report of Examination for the period June, 1978 to December 31, 1974; in another case an actuary resigned after being unable to obtain reliable employee data, see New Orleans Times-Picayune, January 8, 1974.





0A

The number of large public plans increased dramatical ly during the period just prior to the enactment of Social Security legislation in 1935, but before optional coverage was afforded state and local government employees. Over 40 percent of the larger state and local plans have, increased their scope of operations by adding new employee groups to' their coverage. Many new plans have consolidated operations in an attempt to rationalize benefit policy and to gain expertise and efficiencies in plan management, accounting, actuarial, computer, and investment functions. Plan consolidation has resulted in administrative savings and proved beneficial in some cases; but there can be barriers as well as some disadvantages to consolidation.
2. Public pension policy
With notable exception, most governmental jurisdictions, at all levels, have not developed an overall policy framework to serve as a, guide for dealing with public pension issues. Historically, the PERS has taken shape through the addition of a patchwork of laws and pI'ograms, creating complexity and confusion. A unique aspect of the public sector in dealing with pension matters is the reluctance or inability of one legislature or administration to commit its successors to a particular policy or course of action. The absence of a discernible pension. policy may result in the lack of meaningful financial benefit disclosure; the lack of auditing and actuarial standards and reports; inequitable benefit structures; rapidly escalating pension costs resulting from inadequate levels of funding; and complexity and confusion caused by divided pension responsibilities among different jurisdictions and between different agencies or branches of the same jurisdiction.
Some states have formed retirement commissions to enable legrislatures and administrations to deal with pension matters more effectively on a long-range basis. States with retirement commissions are more likely to have catalogued information on the plans operating within their jurisdictions than other states, many of which were found to have little knowledge of the number and types of plans operatingwithin their boundaries.
3. Retirement system administration
While the survey data show that larger plans are generally better managed than smaller plans, in some cases larger plans display the same administrative weaknesses as smaller plans.
With rare exception, the administrative responsibility for large. public plans is vested in a retirement board, whereas plan administration in nearly one-third of all smaller state and local pension plans is carried out under the offices of elected public officials. The effectiveness and sensitivity with which a retirement board can deal with complex administrative and investment issues is dependent on the interest, experience, and abilities of its individual members. As diligent as some administrators have been in fulfilling their duties, the. lack of clear-cut pemlsion policies, direct ion, and control has preven tcd some systems from overcoi n rg cert4a in administrative wen knes,(:s. This has been most apparent in the administration of the dis:ability pr-ograms of some systems-lar'ge and small. Administrative laxity in the disability ar-ea hias forced at least one plan in the past into court ap)oifltod receivership. In other b~enefit areas, administrative p)1oce-







dures have prove& inadequate to prevent certain benefit provisions from resulting in favoritism, other inequities, and -abuse.
Retirement'system audits
In general, public pension plans at all levels of government do not appear to be, operated within the generally accepted finan cial and accounting procedures applicable to private pension plans. The most striking characteristic of public employee retirement system audits is the absence of any uniform or standard auditing practice. Substantial segment of the public employee retirement system is characterized by the absence of any regular or external independent review of plan operations. Nearly one-third of all state and local plans and 37 percent of the larger ones do not provide for annual audits. Some of the largest public pension plans in the country are only audited every four or five years, and they are audited at that time by related governmental agencies rather than an external independent auditor. About 4.6 percent of all state and local plans and 29.1 percent
-of all federal plans are not audited at all. The lack of a regular, independent review of public pension plan financial and actuarial matters caxries an attendant risk of financial miscalculation or abuse.
5. Plan disclosure to participants
In summary, the disclosure practices of public employee retirement systems at the federal, state, and local levels fall considerably short of the high standards set for private pension plans under ERISA.
A wide range of public pension plan disclosure practices exists in the PERS. In many cases plan disclosure to participants is inadequate or nonexistent. In such cases plan participants and beneficiaries (and other interested parties as well) are unable to assess properly plan financial operations, are unappreciative of the true level of pension costs, and are. unaware of conditions leading to benefit losses.
Over one-fifth of aU state and lo cal pension plans do not prepare or supply plan participants with plan descriptions. Less than a majority of the state and local plans and 69 percent of the federal plans make a regular practice of updating and distributing plan descriptions of one, form or another.
Where plan descriptions are furnished automatically or upon request, there exists great variation in the form and usefulness of such materials. Some of the largest plans distribute elaborate and comprehensive plan "booklets". In other cases descriptive "pamphlets" axe so brief as to be of little help to participants seeking answers to practical questions.
A glaring weakness of contributory retirement systems, is the inadequate means of informing participants of their rights to vested benefits or to the return of their own contributions upon termination. About one-half of all federal, State, and local plans. do not automat.1cally furnish employees with statements of their contributions. In over 8 percent of the plans, the participants are unable to receive such statements even upon request.
The results are predictable. Many pension plan membership rolls and trust funds are clogged with the names and accumulated contributions of former members with which the plans have lost total contact. The problem is a multimillion dollar one, and the only way some governmental jurisdictions have acknowledged the problem is by






82

trying periodically to expropriate the unclaimed pension fundsunder state escheat laws. The problem of unclaimed. benefits has been a perennial one for the Federal Civil Service Retirement System. A 1972 General Accounting Office report estimated 338,000 former federal employees aged 62 or over as having potentially unclaimed benefits totaling $26 million. The fact that the federal system retains the contributions of approximately two million former federal employees suggests that the problem of unclaimed benefits will continue for this system as well.
6. Internal Revenue Code qualification
Generally, public pension plan officials have been unmindful of -the tax obligations extending to employees and their pension funds in the event certain qualification requirements of the Internal Revenue Code (e.g. IRC section 401(a)) are not met. The uneven enforcement by' the Internal Revenue Service of Code provisions applicable to public pension plans has undoubtedly contributed to the confusion of public officials as to their obligations in this area. Nearly 58 percent of the state and local and 15 percent of the federal plan administrators prof essed to. be unfamiliar with the application of In temal Revenue Code qualification procedures. Only about 15 percent of the state and local and 50 percent of the federal plans have applied to the IRS for a determination of their qualified status. Because public plans have not been uniformly subjected to the qualification provisions of the Internal Revenue Code in the past, many public plans lack the safeguards inherent in private plans even prior to ERISA.
7. Recordkeeping and other procedures
Early public employee disability and pension programs started out as extensions of city pay-roll operations'. In a few large cities and numerous small ones pension operations have not progressed much beyond this early stage. Most large city and state systems, however, have made substantial progress, many in the last decade or so, in brii ging modern management and computer techniques to bear on their pension operations. Nonetheless, large public employee retirement systems of the multiple employer variety experience some unique adminl-strative and recordkeeping problems of their own. Obtaining and maintaining recorded data relating to, in some cases, thousands of participating employers and several hundred thousand employees present unique problems for multiple employer plans. Multiple emplover plans often lack up-to-date information on the membership and ern j )loyment status of individual employees. As a result such plans m.q-k- not know the number of terminated vested employees and may have trouble estimating the related effect on benefit costs.












CHAPTER D-PERS PARTICIPATION, VESTING, PORTABILITY, AND~
PLAN TERMINATION PROVISIONS
The Employee Retirement'Income Security Act of 1974 (ERISA) extends important pension plan protections to employees in the private sector. These protections include minimum standards relating to pension plan participation and vesting (ERISA Title I, Part 2). The law also provides for termination insurance (ERISA Title IV) which guarantees the payment of vested pension benefits, subject to, certain limitations, when a plan terminates with insufficient assets. Certain provisions under ERISA also permit a modicum of pension. portability. This chapter discusses the extent to which public employees enjoy similar protections under federal, state, and local govern-ment pension plans.

PARTICIPATION (MEMBERSHIP) REQUIREMENTS
Generally, ERISA requires private pension plans to begin accruing benefits for employees who are age 25 or older after they have completed one year of service (employment). A provision requiring employees to''be age 25 and to'complete three years of service may be suitued if the plan has immediate vesting. Also, any plan maintained for employees of an educational organization which is taxexempt under Internal Revenue Code section 501 (a) may use an age 30 and'one year of service' eligibility requirement, if it provides immediate vesting. Employees meeting the minimum age and service r Ie quirements muist be permitted to commence participation at the beginning of the 'next plan year, thus' extending the service period beyond one year or three years, but not -more than six months beyond such period.
It can be seen (Table Dl) that most public employees are in plans with more liberal participation provisions than those required by ERISA. Nearly 60 percent of the total number of state and local government plans covering 97 percent of all active participants have no minimum age requirement. Approximately the same number of plans, covering nearly 84 percent of all active employees, have no minimum, service requirement. The two largest federal plans, the Civil Service Retirement System and the Military Retirement System, both permit immediate participation (i.e. no minimum age or service is required). The most notable exception to the typical "no minimum age") provision is the 21 years of age requirement found in 50 percent of the police and fire plans covering 30 percent of the employees in that category.
Approximately 2 percent of the public employees are in pension Plans having age and service requirements which do not meet ERISA minimum standards. The plans requiring higher ages or greater lengths of service for participation than ERISA amount to between
(83)





Q4

14.2 percent and 17.6 percent of the total number of state and local government plans. Such plans typically have less than 1,000 active lparticip~ants.
Another ERISA age-related eligibility condition restricts the maximuim age limitations which are commonly found in private pension plans. This ERISA rule prohibits a private pension plan from exchuding employees solely because they have attained a particular age. An exception to this rule permits defined benefit and target benefit plans to exclude employees who are hired within 5 years of a plan's normal retirement age (usually age 65).
It can be seen (Table Dl) that only 7.7 percent of the state and local government employees are covered by plans having a maximum age limitation of less than age 60 (which might result in more restricted participation at the older ages than would be permitted under the ERISA rule). For such plans covering general government employees the maximum age limitation is typically age 55. Since the normal retirement age for many of these plans is also age 55, the maximum age limitation is probably not more restrictive than that permitted under the ERISA rule. However, this is not the case for many police and fir'e plans. Over 35 percent of the police and fire plans covering nearly 25 percent of the total employees in that category have a maximum agre limitation of age 35 or less. This age usually. corresponds to the maximum age established by police and fire departments for hiring purposes.
The two largest federal plans covering civilian and military personnel do not have 'a maximum age limitation. However one-third of the federal plans do have a maximum age limitation of 'between age 60 and 64. Four federal government plans have maximum age limitations, between ages 56 and 58, which would not meet the ERISA standards. Three of these four plans cover the nonap propriated fund employees of the U.S. Coast Guard, Navy, and Marine Post Exchanges. The other plan is maintained by the Columbia, South Carolina Farm Credit District.










85.


co cl c) 4= 4= m CD CD CD
c, =c;,c
96 0-4 -4 #-1


co (.0 = (30 u) (D -tr m C) Lim -tr
-cc E S
4 ad i.-4 4= cy; c-i cd

E -:a F = I
va
-v EO 1 LO M oj cn "4

E
Is


cr r,: ci ui vi (6 cD oi
V.4.-4 cn

CL

co r-. cn U-) W m M C*4 M to (D r m
L- V%
f-: oi 06 06 06
01 C-4 m C-4 cn



E cn cJ r zr m rl- .41
E cll C-i (z; c-j 1 6 cli vi lr : od 6 ui
E
.E



CD (D
4


C C) en "t co en vt Lo, oo 00 CD c1l, M
C4 -4,11diC4660i r_: Cd c-i cd Ad
Wo U") Lc) cy),qr qw,;T t %r cn CY)
00

,2=
LA6j 00
m c CD M en mr r- C-4 m CD 00
E wo) -ci C-4 "4
= cl Ln
E cc
uj
cl: 4) Rt M U,) Ln M
C6
r-Z 6od ui ad ri -cr -4 od
0) 'tr M LO Ln -W --r -W ;rmw -cr

LU C>
40 CD =
1-4 -4 "4 -41-4 -4 -4
Uj 4)

>U-3, ..4 00 1. -4 -R U" 4D
w
C. cc
CL
cr

w U-J to -r Lim Ln to c) -4com Lo
cc Cj to
a C --i r.: vi cd
c"
E 04

E
E
4) to Ln C" 00 -cr w
w cy; cli cli r-4 cy; 06 06 P : ca r-: cli 4.6 06
cr cn 0 4m to (n c) M 0 w W-d-m Ln

CD ==84=c m = C) ===
C: CD C) 4= CD CD==
.4 -4

ui
-j 0 M U'Y C ul.-4 um
c'! cvi cwi
r E
C-4

cr
w
M"m
w Cc >, vi evi cvi -o, 6 ui ui Cd
-4 "4 4 C14 Cj C14 04
>
(D
_j

CD eq m 00 Lin" en C* to 00 4m.-, -qr
0; vicr* M* L6 Cj M, vi 0; ui ci r-: cy; C:L
4" cowwwwr- w w ul4, Q U: Lin Q.


0

I f U




En (D U)
E
LAJ a
CA
LAJ t cc
0 it ba
>
00
to 4)-s
DA
U.1 LA- 0 =.E
-:E
E U..

fU >- : W M I 'a >
4>15 CID _C 0- W=
oG = -b-- bO-0.2!.2 0
to
ca
.6.1
L.1 LA- LZ w






86

ERISA generally requires that pension benefits be made available at normal retirement age to persons retiring with five years of service. An exception to this rule is allowed under ERISA when normal retirement ag~e is defined as the later of age 65 or the 10th anniversary of the time the employee commences participation in the plan. While public employee plans are generally free of restrictive maximum agre limitations, with the exception of the police and fire plans noted above, many plans do require more than five years of service (or participation) before benefits become payable regardless of whether participation commences before or after normal ret irerment agye. The service requirement is usually the same as that required for vesting (which is taken up in the next section).
The one aspect of participation in.-w7hich public pension plans fall considerably short of the ERISA minimum standards involves the coverage of part-time employees. ERISA requires that a "year of service" for participation purposes be credited an employee who works at least 1,000 hours during a 12-month period. It can be seen (Table D1) that three-quarters of the state and local plans, covering 37.4 percent of the total membership, do not extend plan participation to employees working 1,000 hours per year. Undoubtedly many of the part-time employees working at least 1,000 hours per year are also excluded from the 16.4 percent of the plans (Tables D1), covering 36.4 percent of the total membership, for which the minimum hours for eligibility are unknown.
Using full-time only coverage as an indicator, plans covering teachers (other than higher education) appear to be the least restrictive in their coverage of part-time workers while police and fire plans are the most restrictive. For plans including part-time employees, 1,040 hours per year is one typical coverage requirement, thus indicating that public employee retirement systems have also given recognition to the ERISA principle that employees who work at least "half-time" should be covered.
Because many plans require a restrictive number of hours service for participation, part-time public employees make up the largest segment of the public sector workforce who remain without pension coverage. As a result, probably more than 50 percent of the public employees working less than full-time lack pension coverage (see also Chapter B).
The coverage of part-time employees is a factor which also relates to the "breadth of coverage" requirements of section 401 (a) (3) of the Internal Revenue Code (as in effect before ERISA) which continue to apply to governmental pension plans seeking to maintain their qualified status on a current basis (see Part Il-Internal Revenue Code for a discussion of how the IRC applies to governmental plans). Generally section 401 (a) (3) requires that for a governmental plan. to, qualify 70 percent or more of all employees, exclusive of short service (five years or less), seasonal (5 months per year or less), and parttime employees (20 hours per week or less,), musi--t be covered or at least 70 percent of the employees (similarly defined) must be eligible for coverage under the plan and 80 percent of such eligibile employees must actively participate.
It appears that few, if any, large public employee retirement systems would have problems meeting this standard, even though participation






87

m-Iay be limited to full-time employees. However, small public pension plans-particularly those maintained at the township level where 45% of the employees work part-time-may have to include some parttime employees, if the Internal Revenue Code qualification require.ments are to be satisfied.

VESTING AND RELATED REQUIREMENTS
Generally, ERZISA requires that the portion of an employee's accrued benefit derived from his own contributions be nonforfeitable (i.e. 10-0 percent vested). The portion of the employee's accrued benefit derived from employer contributions must be vested in accordance with one of three minimum vesting rules:
1. Ten-year service r'ule.-100 percent vesting after 10 years of service; 2. Graded 16-year service rule.-25 percent vesting after 5 years of service; then 5 percent additional vesting for each year of service from year 6 through 10, then 10 percent additional vesting for each year of service from year 11 through year 15, so that the employee is 100 percent vested after 15 years of service; and
3. ul of46-50 percent vesting after 5 years service or, if later, when agre plus service equals 45, such percentage increasing by 10 percent each year until 100 percent is reached; additionally a participant under the Rule of 45 must be 50 percent vested after 10 years of covered service, such percentage increasing by 10 percent for each additional year of covered service, so that the employee is 50 percent vested after 10 years and 100 percent vested after 15 years regardless ~of age.
The vesting provisions in the retirement plans for public employees are compared with ERISA minimum standards in Table D2. Nearly 71 percent of the total plans covering slightly over 20 percent of the total number of state and local government employees have either no vesting or later vesting than that required by ERISA.
Only 2.4 percent of the state and local government employees are in plans with no vesting. Nearly all of the employees without vesting are local policemen and firefighters (26 percent of all such employees) who are usually required to work 20 or 25 years to obtain a pension. The 2.2 million members of the federal uniformed services -remain the largest group of employees, inside or outside government employment, with no pre-retirement vesting of pension benefits.
At the other end of the vesting spectrum, over 38 percent of the total. number of state and local government employees are covered by
-pension plans with vesting after 5 years or less. Less than 4 percent of such employees have immediate vesting, although 36 percent of the teaching staff of public colleges and universities obtain immediate vesting under their defined contribution plans funded through the 'Teachers Insurance and Annuity Association and its affiliate, the 'College Retirement Equities Fund. Several small federal plans are
-also funded through TIAA-CREF and have immediate vesting (see Table VI of Appendix IV). With minor exception the remaining fed,,eral civilian employees' obtain vesting after five years of service.
As is the case in-private pension plans, the most typical public pension plan vesting provision is 100 percent after 10 years of service. Neaiily -30 percent of -the plans covering over 40'percent of the state






88

and local government employees provide straight ten years of service vesting. Slightly less than 9 percent of the state and local plans, covering about*5 percent of the total employees, require a minimum age and service time for vesting. Nearly all such plans fail to meet ERISA's "Rule of 45" vesting provision, Only 8 percent of the total number of state and local government employees are covered by plans (comprising about 20 percent of the total number of plans) utilizing a graded vesting schedule. It, is believed that the graded vesting in most plans would meet the "Graded 15-Year Service Rule" under ERISA.
TABLE D2.-RETIREMENT SYSTEM VESTING PROVISIONS fin percent]
Plans with service vesting
only Plans Total
Plans with age plans
with no 5 yrs 6 to Over and service with Employee category vesting or less 10 yrs 10 yrs vesti ng vesti ng Total
EMPLOYEES
Federal ------------------------------- 43.3 54.7 1.0 .3 .7 56.7 100
State and local total:
State---------------------------------------- 39.7 47.7 10.6 1.9 100.0 100
Local------------------------------ 1. 1 34.3 43.5 18.3 2.6 98.9 100
Police and fire------------------------ 26.0 16.9 22.9 23.5 10.7 74.0 100
Teacher------------------------------------- 41.4 37.8 12.3 8.4 100.0 100
Teachers (higher education)--------------------- 66.8 26.6 4.4 2.2 100.0 100
Total ----------------------------- 2.4 38.3 41.2 13.1 4.9 97.6 100
PLAN S
Federal ------------------------------- 25.6 41.9 14.0 4.7 14.0 74.4 100
State and local total:
State and local government------------- 13. 5 24. 1 34. 7 12. 8 14. 8 86. 5 100
Police and fire------------------------ 72.9 3.2 8.0 8.5 7.3 27.1 100
Teachers (including higher education) -------------- 91. 0 5. 8 1. 8 1.5 100. 0 100
Total---------------------------- 52.8 15.0 14.5 9.0 8.7 47.2 100
Note: Data relates to tables 21 and 22 in app. 1.
The effect of the vesting provisions shown in Table D2 is that :38. percent of the total number of state and local government employees currently have attained a vested right in their accrued pension b~enefits (employer financed port ion).'1 The percentage of active employees with vested rights in locally-administered plans, 32 percent, is less than the percentage of vested employees in state-administered lans, 39 percent, reflecting the more restricted vesting found among local plans (especially those covering policemen and firefighters). The much larger percentage of vested employees in the Federal Civil Service Retirement System, over 60 percent, is reflective of the five year vesting provisions in that plan.
A substantial percentage of the remaining employees in public plans with vesting, but who are not currently vested, will ultimately achlieve vested status. For example, the experience of one large public employee retirement system with ten year vesting shows that a newNly hired employee has the following chance of attaining vested status22 percent if hired at age 25, 48 percent if hired at age 35, and 56
1See table 23 In appendix, I for more detail on the percentage of active public employees who are currently vested.






S89

percent if hired at age 45.2 The employees in. plans requiring more than ten years for vesting have a much smaller chance of achieving 'vested status.
Several factors indicate that excessive. cost does not exist anti should not serve as a barrier to bringing the plans w hich do not now meet ERISA vesting standards into conformance with the "Ten Year of Service" vesting rule. First, nearly all of the employees in such plans already have vesting after 15 or 20 years of service. Second, experience studies. obtained from actuarial reports supplied to the Pension Task Force indicate that the number of employees terminating with ten to 15 (or 20) years of service is typically less than 10 percent of the total number of employees who now terminate employment without vesting. Termination rates are usually much lower in police and fire plans, which are the plans most likely to have restrictive vesting or no vesting, than is the case in plans covering other categories of employees. These first two factors suggest that while a small minority of employees would benefit from the extension of ERISA vesting standards to the plans not now meeting them, the employees benefiting would be those having Substantial peiods of service.
A third f actor which reduces the employer's cost of vesting under public employee retirement systems is the contributory nature of most plans. For example., accumulated employee contributions of, say, .5 percent of pay may be sufficient to "purchase" (depending oii length of service and age at retirement) 20 percent to 30 percent of the total retirement benefits for an employee in a typical plan (see Chapter E).. For an employee with vested benefits who terminates at an age considerably beore normal retirement age, accumulated emplIoyee contributions may "purchase" over 50 percent, and in extreme cases 100 percent, of the employee's vested pension benefit. A study of the cost of vesting commissioned by the Pension Task Force shows that employer pension costs are increased by 3 percent to 6 percent if 10-year vestiifg is extended to noncontributory plans with turnover experience typical of public pension plans.' The contributory nature of public pension plans and the other factors discussed above suggest that the relative cost increase. for most public plans should be less than one-half this level.
As mentioned at the beginning of this section,. ERISA requires 100 perce nt vesting of the -portion of an em-ployees' accrued benefit derived from his own contributions. For defined benefit plans, the portion of the employee's accrued. benefit attributable to employee contributions is 'derived (in accordance with Section 204 (c) of E RISA) by multiplying the employee's accumulated contributions by an actuarial factor .which converts the lump sum accumulation into an annual pension amount. When an 'empDloyee terminates employment and receives, whether voluntarily or otherwise,. a lump sum distribution instead of a deferred pens 'ion benefit based on his own contributions, the lump sum is calculated as, the present value of the annual pension benefit..The lump sum may differ from the accumulated amount of the contributions made by the employee. Although final regulations under ERISA have not been issued in this area, it
2 For example, see the experience for group 3 of. Appendix A, "Estimates of the Cost of Vesting in Pension Plans," by Howard'B. Winkievoss. 'Committee print, committee on Edu!cation and Labor, U.S. House of Representatives, 93d Cong., 1st semS
3See footnote 2, vesting costs for group 3 are shown on p. 18.






90

is likely that the lump sum return of employee contributions will usually have to be accompanied by the payment of interest, at least on a partial basis.
If the above ERISA rules were to apply to public employee retirement systems, over one-half of the plans covering over one-fourth of all state and local government employees would probably fail to meet the minimum standards (see Table 20 in Appendix 1). Nearly 50 percent of the plans covering almost 18 percent of the total number of employees do not pay interest on mandatory employee contributions. Likewise, the Federal Civil Service Retirement System permits the return of employee contributions without interest to persons who separate with at least 5 years of service. Presently, interest is credited at the rate of 3 percent on the contributions which are returned to federal employees who separate with less than five years of service.
Surprisingly, over 12 percent of the total number of local police and firefighters are in plans which do not permit the return of employee contributions upon separation before retirement. Among the different categories of public employees, teachers are the most likely to be covered by plans meeting the ERISA-like interest rule, while police and firefighters are the least likely to be so covered.
Another provision in ERISA prevents the forfeiture of the vested portion of the employee's accrued benefit derived from employer contributions in the event that an employee withdraws his own contributions. An exception to this rule (see ERISA section 203 (a) (3) (D))
-applies to an employee who is less than 50 percent vested. While the vested benefits of such an employee may be forfeited at the time he withdraws his own contributions, the plan must contain a "buy back" provision permitting the employee to fully restore the forfeited benefits upon the repayment of his contribution plus 5 percent interest.
Less than 3 percent of the total -number of state and local government employees are covered by plans containing an ERISA-like provision prohibiting the forfeiture of vested benefits at the time an employee withdraws his own contributions (Table 20 in Appendix I also shows the federal plans to be lacking in this regard). However, 85 percent of the state and local government employees and 100yercent of the federal employees are covered by a "buy back" provision permitting employees. to redeposit witlidra,"m contributions in order to restore prior service credits. Among the different categories of employees, state legislators, 93 percent, are the most likely to be covered by "buy back" provisions while police and firefighters, 62 percent, are the least likely to be so covered.
As discussed in Chapter C, the Pension Task Force survey found many -plans to have relatively few terminated vested employees on their deferred pension rolls. This suggests that the withdrawal of employee contributions by terminated vested participants is a practice that is widespread among public pension p ans. Because accrued benefits related to employer contributions are forfeited in nearly all such ca,-,.es, hundreds of thousands, perhaps millions, of public' employees have lost valuable benefits by withdrawing their own contributions. This unfortunate situation exists because most public'emplovees are not told that they may be forfeiting benefits of much greater value when they elect to withdraw their own contributions.





91

In order to make vesting more attainable and meaningful, ERISA contains certain rules relating to service, "breaks-in-service", and the form and payment of benefits.
Generally, ERISA requires that all of an employee's years of service (employment) with the employer or employers maintaining the plan be credited for vesting purposes (see ERISA section 203(b)). Most public pension plans also follow this rule, although just over 25 percent of all federal, state, and local plans do not credit pre-participation service for vesting purposes (see Table 24 in Appendix I). Because of the generally non-restrictive participation rules under public pension plans (as discussed in the previous section), the provision in 25 percent of the plans defining years of service to be years of participation is not overly restrictive when compared with ERISA.
This is especially so in light of the fact that an exception to ERISA's general rule permits service before the age of 22 to be excluded for purposes of the "10-Year Service" and "Graded 15-Year Service" vesting rules.
Another exception to ERISA's general rule that all service be credited for vesting purposes permits plans to disregard service prior to a break-in-service as defined by the minimum standards. Generally, the break-in-service rules prevent short periods of interrupted employment, whether voluntary or involuntary, from being used to unduly frustrate the achievement of an employee's vested status. The ERISA rule-of-parity prevents a plan from disregarding prior service for a non-vested employee until the 1-year breaks-in-service equal or exceed the prior years of credited service.
Nearly 90 percent of the state-administered retirement systems, like the Federal Civil Service Retirement System, aggregate all service and, thus, do not disregard prior service because of a service break (see Table 24 in Appendix I). In contrast, over 46 percent of the locally-administered plans and 17 percent of the federal plans require that employment be continuous until vesting is achieved. Slightly under 10 percent of all plans disregard prior service for vesting purposes when a break-in-service occurs which is between 1 and 5 years in length.
Generally ERISA requires that deferred vested benefits be payable no later than normal retirement age (see Chapter E). In the case of a plan which provides for early retirement for active participants, a terminated vested participant is entitled to receive his benefits (actuarially reduced) at the early retirement age provided the early retirement service requirements are also met. Considering the fact that deferred vested benefits are payable at 65 or before in all public pension plans, with nearly 75 percent of the employees permitted to receive them at age 60 or earlier, it would appear that the above ERISA requirement is already being met for most public employees (see Tables 21 and 22 in Appendix I).
A final provision of ERISA which relates to vesting is applicable to governmental pension plans which desire to be qualified under the Internal Revenue Code (see Internal Revenue Code section 411(e)). This provision states that the pre-ERISA Internal Revenue Code sections 401(a) (4) and (7) continue to apply to governmental pension plans. Generally, these rules require that contributions and benefits not discriminate in favor of the highly compensated employees over
74-365-78-7





92

the rank and file employees and that the accrued benefits of all employees be vested to the extent funded in the event of a plan termination or complete discontinuance of contributions.
In the past, the Internal Revenue Service has challenged the qualified status of some governmental pension plans for alleged violations of the non-discrimination requirement of I RC section 401 (a) (4). For example, in 1975 the IRS threatened the Missouri PERS with a $5 million tax lien in the event the system did not come into compliance with the Internal Revenue Code.4 The issue involved the plan's earlier vesting for state executive officers and legislators (4-year and 6-year vesting, respectively) as opposed to the 10 years of service required of other state employees for vesting. The earlier vesting of executive officers, legislators, and judges is a practice commonly found in many states other than Missouri. The manner in which the IRS will resolve this and other issues reg-ardingr the application and enforcement of Internal Revenue Code provisions relating to governmental pension plans remains to be seen. See Part II of this report for a discussion of the current position of the IRS in regard to these and other matters affecting governmental plans. The current and past confusion suggrests that a legislative remedy may be necessary to clarify public policy in this area.

PORTABILITY PROVISIONS
The preservation of public employee pension credits takes on several different forms. The complete transfer of pension credit's through par-. ticipation in Social Security is the most important portability pr otection for public employees, as it is for employees in the private Sector. However, over 30 percent of the employees of State and local governments and 55 percent of the employees of the federal government remain outside the scope of Social Security coverage (see C'hapter B).
With regard to the preservation of the pension credits earned by public employees under their retirement systems, vesting is presently the most important means of such benefit protection. However, as discussed in the previous section, even vesting is not universally found among public employee retirement systems.
While Social Security coverage and vesting serve as the. only means of retirement benefit protection for an employee who leaves public employment for a job in private industry, other means of pension portability exist for employees who transfer to employment within the governmental sector. These other forms of portability are usually more valuable to the employee than is vesting alone. This is because, for purposes of the pension benefit computation, the service credits which are transferred in accordance with the portability scheme are ordinarily used in conjunction with the employee's final average wvages at retirement, and not the generally lower wages at the time of transfer.
It can be seen (Table D3) that nearly 82 percent of all state and local government employees are covered by plans permitting some form of inter-governmental pension plan portability. Only 8 percent of the plans covering 3 percent of the total number of state and local gYoverninent employees give automatic credit to employees for service rendered with other governmental employers without requiring the eml)loyee or former employer to contribute or reciprocate in some manner.
4'"Pensions and Investments," Mar. 31, 1975.