INCOMPLETELY VESTED PENSION PLANS AND LABOR MOBILITY Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY EDWARD JOHN FLANAGAN 1969 ‘ VJ LIP r t? Q .~( .- ‘ - - ' t 5 l—fl i We“: "4-:- 471» at “A *'W , . 1 1 l I This is to certify that the thesis entitled Incompletely Vested Pension Plans and Labor Mobility presented by Edward J. Flanagan has been accepted towards fulfillment of the requirements for Ph.D. degree in ECONOMICS Date 0-169 INCOMPLETELY VESTED PENSION PLANS AND LABOR MOBILITY By Edward John Flanagan The freedom of labor to enter and to leave employ- ment is one of the fundamental requisites for economic efficiency. Received theory identifies this freedom as the necessary condition facilitating response to the wage signals of the marketplace. This study deals with the impact of incompletely vested pension plans upon the mobility of labor. The presumption that pension plans impede the mobility of workers is widely accepted. A number of investigations have been undertaken to test the validity of this presumption. In general, these investigations have suggested that pension plans have not had a signifi- cant impact upon worker decisions to voluntarily leave »Jobs. These investigations, however, did not involve tests for ekistence of a relationship between participant's non- vested retirement losses and mobility. Rather the tests were conducted to measure the strength of impediments to labor mobility inherent in non-vested pension plans per se. Edward John Flanagan The hypothesis this study is designed to test is that the significant impediment to labor mobility is the income value of retirement benefits lost by non-vested partici— pants through separation, rather than the fact of their merely having non-vested status. The study identifies those factors determining the value of participants' retire— ment benefits, and then tests for a relationship between participants' non-vested retirement values and impediment to mobility. Quit-rate data were collected and analyzed for two samples of university faculty (one sample composed at faculty at schools having deferred vesting plans; the other composed of faculty at schools having vested plans). The major conclusion of the study is that, while no evidence of impediments to mobility from pension plans, per se, was found, quit-rates of non-vested faculty in the deferred vested plan diminish as the strength of the factors influencing the value of the income stream of benefits lost increases. INCOMPLETELY VESTED PENSION PLANS AND LABOR MOBILITY By Edward John Flanagan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1969 ACKNOWLEDGMENTS I wish to take this opportunity to thank Dr. Charles Larrowe and Dr. Bruce Allen for their constructive criticisms of earlier drafts. Special thanks to Dr. John Henderson for his consul throughout the writing under the difficult circumstances of having to direct the disserta- tion by mail. My colleague at Bradley, Dr. Robin Linstromberg, has my grateful acknowledgment for offering numerous worthwhile suggestions throughout the entire writing period. Also, for her cheerful and capable assistance in typing earlier drafts, I wish to thank Mrs. Lynn Park. Finally, my wife deserves special considera— tion for shouldering much of the male chores encountered in raising a family to allow me the time for research and writing. ii TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES Chapter I. II. III. IV. INTRODUCTION Statement of the Problem. . . . Hypothesis . . . . . . . Purpose of the Study . . . . . . THE THEORETICAL FRAMEWORK . . . . Efficiency Conditions. . . Pensions and the Efficiency Condition Incomplete Vesting and Economic Efficiency. Non- vested Pensions and Labor Mobility PENSION-MOBILITY STUDIES Opinion Studies. . Quit- Rates and Non— vested Funds Concluding Remarks. . . . EMPIRICAL TEST OF THE PENSION-MOBILITY HYPOTHESIS . . . . . Sample Selection Non- vested Sample Vested Sample. Comparison of the Samples Summary of Results. SUGGESTIONS FOR FURTHER STUDY. Plan Differences and Their Impact Upon Mobility of Non-vested Participants. . Financing . . . . . . . . Deferred Vesting. Retirement Formulas. . Mobility Implication of Vesting Conclusion . BIBLIOGRAPHY iii Page ii iv [.1 O\C\U'l l3 16 21 27 28 A7 80 80 82 83 86 89 92 Table 10. ll. 12. LIST OF TABLES Contribution and Benefit Data for a Hypothetical Pension Plan. Opinions of Union Stress on Fringe Benefits Compared to Wage Increases . . . . . Relative Preferences for Selected Fringe Benefits Amongst Steelworkers and Non-Steelworkers, Pittsburgh, May-June 1959. . . . . . Importance of Delayed Vesting in a Decision to Accept an Outside Offer . . . Voluntary Separation Rates, According to Length of Covered Service, in Vested and Non-vested Institutions of Higher Education, 1959 . . . . . . Faculty Sensitivity Ratios, According to Length of Covered Service, in Colleges and Universities by Type of Control, 1959. . . . . . . . . . . Illustration of Monthly Survivors Benefits. Non-vested Credit Value Per Year to Age Sixty Higher Offer Necessary at Vesting to Equal $1,000 Higher Offer During Various Non-vested Periods . . . . . Average Compensation, Full—Time Faculty. Voluntary Separation Rates, Non-vested Sample vs. Vested Sample Voluntary Separation Rates During the Non-vested Period . . . . iv Page 19 3O 3O 37 “3 A5 55 58 61 65 68 68 Table Page 13. Voluntary Separation Rates by Year of Participation. . . . . . . . . . 69 1A. Age at Entrance to Plan of Departing Non-vested Participants . . . . . . 73 CHAPTER I INTRODUCTION Freedom of labor to enter and to leave employment is one of the fundamental requisites for economic effi- ciency. Received theory identifies this freedom as the necessary condition facilitating response to the wage sig— nals of the marketplace. For some time it has been recognized that certain practices of unions, professional associations, and em- ployers have a restraining effect upon the ability of physically and mentally qualified laborers to enter into a variety of occupations. But until recently, factors inhibiting exit from Jobs have received little attention. As a result of the rapid extension of worker coverage in pension plans, however, attention has begun to focus upon the possible impact of these plans upon the freedom of workers to give up Jobs covered by them.1 lAlthough information concerning the number of work- ers covered by pension plans prior to World War II is not available, the Bureau of Labor Statistics has estimated that only two per cent of the private plans in existence in 1960 were established prior to l9A0. See, U. S. De- partment of Labor, Bureau of Labor Statistics, Labor Mo- bility and Private Pension Plans, Bulletin Number 1307 (June, 196A), p. A. The Secretary of Labor has estimated l The primary purpose of a private pension plan is to provide retirement income to employees. But it has been suggested that a major consideration for employers when contemplating establishment of pension plans has been re- duction of the amount of labor turnover.2 Almost all plans require that employees be in the service of the employer for a specified number of years before they obtain the right to a retirement benefit. In the majority of plans, the employee, upon fulfilling enu- merated requirements (usually stipulated periods of ser— vice and/or attainment of a stated minimum age), can leave the service of the employer and maintain the right to re- ceive accrued pension benefits upon retirement. These plans fall into the category of "deferred vested" pension plans. Other plans require employees to be in the service of the employer at the time of retirement if they are to receive retirement benefits. These are referred to as non-vested pension plans. All employees in non-vested plans, and non-vested employees in deferred-vested plans, combined private pension plan and profit-sharing plan coverage at 25,000,000 workers in 1966, increasing to 3A,000,000 by 1970. See, U. S. Congress, Joint Economic Committee, Private Pension Plans, Hearings, before a sub- ‘ committee on Fiscal Policy, Part 2, 9th Cong., 2d sess., 1966, p. 363. 2Merton C. Bernstein, The Future of Private Pen- sions (London: The Free Press of Glencoe, 196A), p. 11. lose accrued retirement benefits upon voluntary separation from the service of their employer.3 The presumption that pension plans impede the mobil- ity of workers is widely accepted. Texts that include coverage of pension plans usually accept the hypothesis that plans have a restraining influence upon voluntary separation by employees.“ The Department of Labor, al- though expressing concern over the loss of retirement in- come by those employees who separate, also has taken no- tice of the possible impact upon labor mobility of private pension plans.5 A number of investigations have been undertaken to determine the validity of the hypothesis that pension plans impede labor mobility.6 In a study of college and university faculty quit-rates, Lurie observed that faculty belonging to plans that vested benefits after five years 3Some plans make a distinction between voluntary and non—voluntary separation. For example, some plans stipulate that participants may vest in five years if in- voluntarily separated, but may vest only after ten years if separation is voluntary. 14For example: Neil W. Chamberlain, The Labor Sec- tor (New York: McGraw-Hill Book Company, 1965), pp. 606- 60—g Allan M. Carter and F. Ray Marshall, Labor Economics (Homewood, Illinois: Richard D. Irwin, Inc., , p. 556; Richard A. Lester, Economics of Labor (2d ed.; New York: Macmillan Co., 1965), p. 350. 5U. s. Department of Labor, Labor Mobility . . ., p. A. 6Since the following studies will be the subject of extensive review and criticism in Chapter III, below, they are simply identified here. of service were more mobile than were faculty belonging to plans that vested benefits prior to five years of ser- 7 Parnes found no significant difference in atti- vice. tudes of plant workers toward mobility when he compared attitudes of workers in a firm possessing a non-vested plan with those of their counterparts in a neighboring firm which had no pension plan.8 Arthur Ross, in a study of labor turnover in manufacturing from 1910 through 1956, observed a modest reduction in quit-rates in the post World War II period.9 However, Ross concluded that it was most unlikely that pension plans were a cause of the quit-rate decrease.l0 Harold Rubin, through an extensive mail questionnaire, invited the opinions of personnel managers for public institutions concerning the effect of their non-vested pension plans on separation and hiring ll of personnel. Respondents, in general, thought their 7Melvin Lurie, "The Effect of Non-Vested Pensions on Mobility: A Study of the Higher Education Industry," Industrial and Labor Relations Review (January, 1965), pp. 22u—237. 8Herbert S. Parnes, "Workers Attitudes to Job Chang- ing: The Effect of Private Pension Plans," in The Reluc- tant Job Changer, ed. by Gladys L. Palmer (Philadelphia: Ugiversity of Pennsylvania Press, 1962), pp. A5—80, 182- 1 9. 9Arthur M. Ross, "Do We Have a New Industrial Feu- dalism," American Economic Review (December, 1958), pp. 10 Ibid., p. 915. llHarold Rubin, Pensions and Employee Mobility in the Public Service (New York: The Twentieth Century Fund, 1965). plans were an insignificant deterrent to voluntary separa- tions; although there was strong evidence that many felt that the non-vesting element in their pension plans was a significant factor-in the refusal of Job candidates to accept positions at their institutions. Thus, in general, investigators have concluded that pension plans have not significantly restrained employees from voluntarily leaving Jobs. Statement of the Problem The value of a pension plan participant's retirement benefit in almost all pension plans is directly related to the length of service and/or earnings of the participant while in the service of the employer. Thus, variation exists in the amount of benefits received at retirement by participants within the same plan. Also, at any point in time, participants having non-vested status within the same plan will have different lengths of service and dif- ferent earnings. Since retirement benefits are related to service and/or earnings, non-vested participants face dif- ferent retirement benefit losses upon separation. Further- more, such factors as participant age, proximity to vest- ing, etc., affect the value of a non-vested participant's retirement benefit loss upon separation. Thus, assuming that retirement benefit losses influence the strength of the impediment to mobility, participants within the same plan are subject to different pension mobility restraints. Hypothesis In the studies mentioned earlier, investigators did not test for a relationship between participant's non- vested retirement benefit losses and mobility. Rather the tests were conducted basically to measure the strength of impediment to mobility provided by pension plans per se. The hypothesis this study is designed to test is that the significant impediment to labor mobility does not lie in the pension plan per se, but is to be found, rather, in the income value of retirement benefits lost through separation. Purpose of the Study It is the purpose of this study to identify the fac— tors determining the value of participant retirement bene- fits, and to test for a relationship between participant non-vested retirement values and impediment to mobility. Using established tools of economic analysis to de- velop a model of the pension mobility hypothesis, quit— rate data collected from two samples of university faculty (one sample composed of schools using deferred vesting plans; the other composed of schools using vested plans), are ex- amined and analyzed. In Chapter II, the pension-labor mobility hypothesis is developed. The analysis depicts benefits received by pension plan participants at retirement as deferred income. It is shown that where pension plans fully and immediately vest employer contributions, such contributions are correctly considered as deferred income in lieu of current wages and, as such, impart no impediment to labor mobility. However, it is further shown that under deferred-vested and non-vested plans, employers' contributions cannot be considered as payment in lieu of current wages and none vested participants are subject to retirement benefit losses upon separation. It is the value of the loss of retirement benefits incurred by the non-vested participant uponseparation that is the ingredient within pension plans that giVes rise to the pension-mobility hypothesis. The balance of the chapter identifies those factors which determine the value of retirement income of non-vested participants. The literature dealing with pension plans is examined in Chapter III. Investigators have used two general ap- proaches to test the effect of plans on labor mobility. Under the first approach, investigators elicited the opin- ions of employees and employers in situations where incom- pletely vested plans are used. The second method involves measurement of the effect of plans on participant separa- tion rates. The procedure, data, and results of each study are examined and critically evaluated. In Chapter IV, the pension-mobility hypothesis de- veloped in Chapter II is tested by comparing the quit- rates of a sample of faculty belonging to the State Universities Retirement System of Illinois as against the quit-rates of a sample of fully-vested faculty at other mid-western universities. The test period included the academic years l96A-65, 1965-66, and 1966—67, when the Illinois plan vested benefits upon completion of ten years of service. Although the quit-rates were higher for the non-vested faculty sample as compared to the vested sample quit-rates, the trend of quit—rates by year of service (a surrogate for non-vested income value in the non-vested sample), decreased more rapidly for the non-vested sample than for the vested. Thus, the test supported the pension- mobility hypothesis developed in Chapter II. Conclusions and recommendations for further research are covered in Chapter V. Because of the complexity and diversity of pension plans, it is suggested that the ef- fect on the mobility of participants in one plan, due to the loss of retirement benefits incurred upon separation, cannot be projected to non—vested participants in all in- completely vested plans. Research on the possible differ- ing effect on mobility of various types of contribution, vesting, and benefit formula provisions among incompletely vested plans is suggested. Of particular importance is the need for research on the effect on the mobility of vested participants that might occur as a result of peri- odic improvements in benefit formulas. CHAPTER II THE THEORETICAL FRAMEWORK A number of studies have been made to test the hypo- thesis that the lack of immediate vesting of employers' contributions to employee's retirement funds inhibits labor mobility.l These studies have centered on two is- sues: (1) the retirement losses incurred if a laborer loses the non—vested portion of his pension, and (2) the extent to which these pension losses impede mobility. Although some of the studies mention the impact of non- vested pensions on the allocation of resources, none has attempted to explain how this impact prevents society's attainment of maximum economic efficiency. It is the pur- pose of this chapter to explicate the effect of incomplete vested pension funds on economic efficiency. Since the possible effect of incomplete or non- vested funds falls on the allocation of the labor resource between firms having separate pension systems, the initial y ’H lParnes, "Workers Attitudes to Job Changing . . Chapter III, Ross, "Do We Have a New Industrial Feudalism, Do 903; Lurie, "The Effect of Non—Vested Pensions on MObility . . .," pp. 224— 2375 Rubin, Pensions and Employee MObility in the Public Service. lO analysis will be limited to the efficient allocation of the labor resource between such firms. Resources will be considered as being efficiently allocated if, at any point in time, the value of output that conforms to the commun- ity's preferences cannot be increased by the reallocation of existing resources. The analysis will begin with the statement of the conditions of maximum economic efficiency in the alloca- tion of resources and proceed to analysis of the effect of incomplete and non-vested pension plans on the attain— ment of maximum economic efficiency. Efficiency Conditions Assumptions: 1. Perfectly competitive market conditions.3 2. Two firms, I and II. 3. One product, 5. A. One input, labor, a (a labor unit). Utilizing the marginal analysis, the condition of optimal resource allocation is as follows: 2Bela Balassa, "Success Criteria for Economic Sys- tems," in Comparative Economic Systems, ed. by Morris Bornstein (Homewood, Illinois: Richard D. Irwin, Inc., 1965), p. 5. The use of the static efficiency measure of Optimal resource allocation is traditional in economics. This definition capsulizes the meaning of welfare maximi- zation in pure welfare theory. 3This assumption encompasses all of the conditions 0f perfect competition plus the institutions of a market sYstem. 11 I _ II (1) VMPa - VMPa X X I II P or MPP a x Px = MPPa where VMPaxI is the value of the output attributed to the last unit of labor employed in the production of x in firm II. A test of the optimality of equation (1) can be made by assuming the following condition: I II (2) VMPa > VMPa X X I II or MPPa Px > MPPa - PX In statement (2), the value of the output attributed to a unit of labor in firm II in the production of g is less than in firm I. Therefore, the community's output of 5 would be increased through successive transfers of a from firm II to firm I until the change in total output of x attributed to the last unit of a employed by firm I, (MPPaI), is equal to the MPPaII. When this occurs, condi- tion (1) is attained which is a resource allocation opti— mum in the sense that the resource a is allocated between two firms producing the same product, in such a way as to maximize Eli value in production. By removing assumptions (2) and (3) and expanding the analysis to firms in separate industries, the following inter-firm resource allocation condition for optimum effi- ciency holds: 12 I _ III (3) VMPa - VMPa X y I where VMPax is the value of the output attributed to the last unit of labor employed in the production of x in firm I and, VMPayIII is the value of the output attributed to the last unit of labor employed in the production of y in firm III. From equation (1), the value of the marginal product of g in firm I producing product 5 is equal to the value of the marginal product of a in firm II producing product x. Therefore, the value of the marginal product of a,in the production of x in firm II must be equal to the value of the marginal product of a in the production of y in firm III when a is efficiently allocated. This allocation condition can be expanded to include all firms in all industries that use resource a as follows: I _ II _ III = = n (u) VMPa — VMPa — VMPa ... VMPa x x y z where g_represents all products that use a as an input and n represents all firms using resource a. If an inequality exists between the value of output attributed to the mar- ginal unit of g between firms, then the maximum value of output would be increased by transferring a from firms where QLE contributive value in production is greater. The stimulant that motivates the transference of 3 can be ascertained from the following optimal condition: 13 I _ II _ III = = n - (5) VMPa - VMPa - VMPa ... VMPa — w x x y z where Wa is the remuneration received by a for its produc- tion contribution. This remuneration is equal to the present money value of the production contribution of the labor unit, which is in turn equal to the labor units' highest opportunity cost in present money value terms. A behavioral condition of the market system that is implicit in condition (5) is that labor will attempt to maximize the present value of its remuneration. Given this condi- tion, concomitant with the other conditions of a perfectly competitive market system, any change in the VMPa in any firm will result in a change in the Wa for that firm and cause a movement of a from or toward that firm dependent on whether the VMPa to the firm has risen or fallen. Thus, a in quest of higher remuneration will move from lower to higher paying firms and in this manner the efficiency con- dition of equation (5) is approached. Pensions and the Efficiency Conditions One of the assumptions of perfect competition must be made explicit before analyzing the affect of incomplete and non-vested pension plans on the efficient allocation of resources. Perfect competition assumes awareness of market conditions on the part of buyers and sellers. Kenneth Boulding describes this awareness assumption as follows: 1A A competitive market is also one in which the buyers and sellers are in close contact. This means that there must be knowledge on the part of each buyer and seller of the prices at which transactions are being carried on, and on the prices at which other buyers and sellers are willing to buy and sell. It means also that there must be opportunity to take advantage of that knowledge. In the context of this study, this assumption means that a laborer is aware of the present value of his remun- eration and is also aware of his opportunity income else- where. Further, the condition assumes that a laborer is not impeded from taking advantage of his best opportunity, i.e., if his opportunity income is higher in other employ- ment, he must be free to move to such.employment. The opportunity remuneration to labor may be re- ceived in a variety of forms, of which present money in— come is the primary form in this nation. However, as noted above, millions of workers accept a'portion of their current remuneration in the form of pension credits in lieu of current cash payment for their services. When the pension credits are fully and immediately vested, title to the entire employer's current contribution passes immedi- ately from the control of the employer and into the em- ployee's account. From the employee's viewpoint, gaining immediate title to the employer's current pension contri- bution permits him to terminate his employment without l4Kenneth E. Boulding, Economic Analysis (3rd ed.; New York: Harper & Brothers, 1955), p. A6. l5 losing title to the employer's contribution. Mobility of labor is not hindered because of the full and immediate transfer of the employer's current pension contribution to the employee's ownership. An accurate measurement of the current remuneration value of the employer's contribu- tion can be made. Therefore, the necessary condition for economic efficiency of buyer's and seller's having know- ledge of the prices of transactions is not obstructed by a fully and immediately vested pension plan. Thus, the maximum efficiency condition as noted in equation (5) holds since, given knowledge by buyers and sellers of the prices of transactions, the employer's current contribu- tion to the employee's pension fund must be equal to the present cash remuneration given up by the employee in lieu of pension credits.5 Rewriting equation (5) to include fully and immedi- ately vested pension credits as a part of labor remunera- tion, the maximum efficiency condition is as follows: I I I _ I _ II _ III _ a - CVRa - VMPa — VMPa _ x x y ll 0 O + O < (6) VMPa n _ n VMPaz - Wa where: CCaI = the current cash remuneration to labor, a, in firm I producing product x. 5Competition among laborers and among firms will assure this equality. l6 CV3I = the current value of the current employer's contribution to the employee's pension fund CVR I = the current value of the employee's total current remuneration As was noted previously, this is an equilibrium con- dition and any increase in the VMPa in any firm, including firm I, will cause a movement of labor from other firms to that firm, generating a new equilibrium efficiency posi- tion symbolized by equation (6). Incomplete Vesting and Economic Efficiency An incompletely vested pension plan can be defined as one in which the employee does not receive full and immediate title to an employer's current pension contri- bution in lieu of current cash. Because of the variety of forms that incomplete vesting may take, the foregoing definition permits application of results of the follow- ing analysis of the effect of one such plan on economic efficiency to all incompletely vested plans. The analysis begins with the following illustration: Assume the following: 1. A $500 employer contribution is deposited at the end of each period of covered employment into an employee's retirement fund. 2. The employer's contribution is fully and immediately vested. l7 3. The pension credits earn A% interest, com- pounded periodically. A. The employee's current cash remuneration is $10,000 received at the end of the period. Now, assume that the pension plan, although funded by the employer, is completely non-vested until the end of the twentieth year. When pensions are not fully and immediately vested, complete title of the employer's con- tributions does not immediately pass to the employee's pension fund. In fact, from the employee's viewpoint, the pension contract can be viewed as a firm commitment by the employer to contribute to the employee's retire- ment income stream an amount of money whose present value at the end of the twentieth year is $1A,889. If the em- ployee leaves his position any time prior to the close of the twentieth year, the present value of the retirement income stream is zero. Therefore, the nonrvested portion of an employer's current contribution to an employee's pension fund cannot be precisely valued as is the case with vested pension credits. This is not to say that non- vested contributions to a pension plan have no value to an employee. But, the non-vested credits have a measur- able value only when the employee has fulfilled specific covenants stipulated in the pension contract. In this example the stipulation is twenty years employment. When covenants such as these are fulfilled, non-vested credits 18 become vested and assume a measurable present value. It is the possibility of transferring non—vested credits to vested credits at the vesting date (or at retirement) which gives the non-vested credits value. For without it, the value of non—vested credits to the employee would be zero, and he would suffer no loss upon terminating his employment. It is the incapability of full and immediate transfer of non-vested credits to vested credits that causes non-vested credits to have a different current re- munerative value than vested credits. Therefore, refer- ring to Table 1, when the plan is completely vested, the employee is continually capable of measuring precisely the employer's current contribution, $500 at the end of each year. The summation of the employer's current pension contribution of $500 and the assumed current cash remun- eration of $10,000 yields a current value of the employ- ee's remuneration of $10,500 which is equal for all em- ployees. This equality is a necessary condition for economic efficiency as noted in equation (6). If, however, the assumption is made that the pension plan is completely non-vested until the end of the twen- tieth period of covered employment at which time it fully vests, then an employee who reaches full vestment would have a current value of remuneration in that time period equal to $1A,889 plus his current cash remuneration. Em- ployees prior to the full vesting period would have a l9 000.0H 000 000.0H 000 00 000.00 0H0 000.0 000 00 000.0H as: 000.0 000 0H 000.HH 00: 000.0 000 AH 0H0.0H 000 000.0 000 0H 0H0.0H 000 000.s 000 0H 00H.0 000 000.5 000 0H 0H0.0 H00 000.0 000 me 0H0.s 0am 000.0 000 NH 005.0 0:0 000.0 000 AH 000.0 Hem 000.0 000 0H H00.0 30H 000.0 000 0 s00.0 000 000.0 000 0 000.0 00H 000.0 000 a 0H0.0 00H 000.0 000 0 00e.m 00 000.0 000 0 00H.m 00 000.0 000 0 H00.H H0 000.H 000 0 0N0.H 0m 0 000.H 000 m 000 0 000 0 0000 H oEoocH pcmsohfiuom mono c soapsnfihucoo .QOfipsnfinucoo mo osam> ow m mew o>ap0a3§do pconhso newwnmoupcm pcomonm an H o n.9oonmem m.pomoadem .cwHQ cofimcod awoapmnpOst m pom dump ufimoson one soapsnfihpcooll.a mqm p (PVVa) . i where: CVR II= the current value of the employee's total a current remuneration with the new firm CVR I a = the current value of the employee's total current remuneration with the present firm p (PVv )I = the employee's present value of the a non—vested pension credits _ -n 1 (l + i) = the present value of $1.00 per 1 period recieved at end of each period. If, p (PvV )I = $10,000, i = 4%, and n = 30, we have: a. CVR II — CVR I > $10 000 a l ' (1 + '0“) “30 = $578 a a ’ .OA Given condition (7), the employee would have to be of— fered a new salary where the current value exceeds his pres- ent salary by $578, assuming the differential will remain 25 constant to retirement. If the differential increases over time, then the necessary initial differential would be less. The required differential is the measure of the monopsony power of the employee's present employer. A change in any of the other assumptions would alter the size of the necessary differential. Thus, 1. An increase in p (PVva) will result in an in- crease in the required differential. 2. An increase in i will result in a decrease in the required differential. 3. An increase in n will result in a decrease in the required differential. When non—vested credits vest prior to the employee's retirement, an additional consideration is posed. Al- though an employee will benefit from accepting a new posi- tion, (condition 7), the employee will increase his gain by terminating after vesting takes place if the following condition exists: , \ AV (1 + i)5’l X ‘ i (8) R > AV - 1 - (1 + i)”n i L 0 the increase in remuneration per year over his present remuneration after vesting where: R X = the present value of pension credits at the time of vesting AV = the difference in yearly remuneration between a current job offer and his current employment .0 ll :5 ll retirement H. II 26 the rate of interest the number of years until vesting the number of years after vesting until Let: X = $10,000, AV = $1,000, q = 2 years, n = 28 years, and i = A%, we have: / x 2 $1 000 (1 + .OA) -1 $10,000 - ’ .0” R > $1,000 — 28 = $523 l-(1+.0A)‘ .OA \ / Thus, if the employee believes that he will receive an offer at the time of vesting that is in excess of $523 per year over his present job, then he should not accept at the present time an offer paying $1,000 per year more than what he is receiving in his present job. The isolation of influential variables is necessary to test the mobility hypothesis of the effects of vesting and non-vesting of pension funds. The variables chosen for this study were selected after an analysis of other studies concerned with the pension mobility hypothesis. Therefore, the impediment to labor mobility increases 1. the higher the present value of a non-vested pension fund at the time of full vesting. 2. the shorter the necessary time to full vesting, and 3. the shorter the work—life expectancy of the employee. CHAPTER III PENSION-MOBILITY STUDIES Although many economists have suggested that non- vested pension credits inhibit the mobility of labor, rel- atively few authors have attempted to verify the hypothe- sis. Probably the major reason for the limited number of studies in this area is that only recently has a high percentage of the labor force gained pension plan cover- age. Thus, the possibility of non—vested funds impeding the mobility of a significant percentage of the labor force is a relatively new problem. The newness of pension- mobility studies attests to the validity of this explana- tion. If the trend of expanding coverage and benefits persists without significant liberalization of vesting provisions, it can be assumed that more surveys will be forthcoming.1 Two conditions are necessary for validation of the pension-mobility hypothesis. First, workers must be 1U. S. Department of Labor, Labor Mobility . . ., Chart 1, p. 5. In 19A6, less than four million workers (approximately seven per cent of the civilian labor force) were covered by private pension plans. By 1960, the num— ber of covered workers had increased to about 15 million (approximately 23 per cent of the civilian labor force). 27 28 aware of the value of their non-vested credits, and sec- ondly, the knowledge of the potential loss must tie them to their current employment. If the first condition is not satisfied, the second is not relevant. Accordingly, studies concerned with ascertaining the value that em— ployees assign to compensation in the form of pension credits, while not directed specifically at testing the pension-mobility hypothesis, do yield insight with re- spect to the hypothesis' validity. When investigators have attempted to test the pension-mobility hypothesis, two general approaches have been utilized. First, opinions have been elicited from employees, and management personnel, who might possibly be affected by non-vested pension funds. For example, employees covered by non-vested plans have been asked if loss of pension credits would be a significant deterrent to accepting another job. The second approach has been a comparison of quit-rates for employees with non-vested plans as against quit-rates of those employees where such plans do not exist. If the quit-rates for non-vested plans are significantly lower than for employees with vested plans, the pension-mobility hypothesis is verified. Opinion Studies Wagner and Bakerman obtained workers' opinions, con- cerning the value of various types of compensation, by interviewing 300 members of the United Steelworkers 29 (AFL—CIO), and 300 members of other unions in the Pitts- burgh area, May-June, 1959.2 The age distribution of respondents was selected to conform to the age distri- bution of all workers employed in the primary and fabri- cated metals industry. The first question asked was: "Would an increase in fringe benefits satisfy you as much as a direct wage increase if it amounts to the same in dollars and cents?" Of the total number of responses, 91.0 per cent of the steelworkers and 79.7 per cent of the non-steelworkers were positive.3 For the second question: "Does your union stress fringe benefits too much compared with wage increases, or not enough?", the responses were as depicted in Table 2. Finally, the workers were asked to rank four bene- fits which they considered most important. After a simple weighting of choices, Table 3 depicts the results.Ll As indicated in Table 3, pensions ranked first among fringe benefits by weighted percentage in both samples. Also, although not depicted in Table 3, in both 2Ludwig A. Wagner and Theodore Bakerman, "Wage Earners' Opinions of Insurance Fringe Benefits," The Journal of Insurance (June, 1960), pp. 17-28. 3All statistics quoted in the text are from Wagner and Bakerman, "Wage Earners' Opinions . . .," pp. 19-22. “Weighting was as follows: four for first, three for second, two for third, and one for fourth. 30 TABLE 2.--Opinions of union stress on fringe benefits compared to wage increases. Opinions of Steelworkers Non—steelworkers Union Stress Number Per cent... Number ...Per cent Too much 7 2.3 17 5.7 Not enough 1A7 A9.0 112 37.3 Just right 1A3 A7.7 167 55.7 No answer __3 1.0 __A __1L3 Total 300 100.0 300 100.0 Source: Adapted from Wagner and Bakerman, "Wage Earners' Opinions . . .," p. 20. TABLE 3.-~Re1ative preferences for selected fringe benefits amongst steelworkers and non-steelworkers, Pittsburgh, May-June, 1959. Fringe Benefits Steelworkers Non—steelworkers PensiOns 32.9 33.1 Group life and health 27.0 25.7 Group major medical 12.0 11.6 Vacations 6.5 8.A Separation pay ' 1.0 2.7 Supplementary Unemployment 11.6 9.2 Guaranteed annual wage 9.0 __2;§ Total 100.0 100.0 Source: Wagner and Bakerman, "Wage Earners' Opin- ions . . .," p. 22. 31 samples, pensions were ranked first by a weighted percent- age for all age groups. Pensions were ranked first by 55 per cent of the steelworkers, and by 58 per cent of the non-steelworkers. In a follow-up interview of 50 steelworkers in August, 1959, at the time of the steelworkers' strike, 66.0 per cent ranked pensions first among the seven sug- gested fringe benefits. At first glance, Wagner and Bakerman's findings strongly suggest that workers prefer compensation in- creases in the form of fringe benefits rather than wages. However, as with all surveys, there is the possibility that the sample was biased. As Wagner and Bakerman them— selves indicated, there had been a great deal of discus- sion in the Pittsburgh press, prior to and during the interview period, concerning the inflationary affects of wage increases. Therefore, it is possible that the re- spondents were convinced that increases in fringe bene- fits would be less inflationary than wage increases, and that the benefit obtained from wage increases would be reduced by rising prices. However, although one cannot generalize from the study, that all workers favor fringe benefit increases over wage increases, Wagner and Baker- man's findings cast doubt on the opposite belief that workers overwhelmingly prefer wage increases to fringe benefit increases. It seems that worker preference may 32 depend upon a number of factors of which inflation is one such factor. Although expressing surprise that the majority of respondents selected pensions as the most important fringe benefit, Wagner and Bakerman did not offer an explanation for the choice. Normally, one might expect benefits such as a guaranteed annual wage or supplementary unemployment compensation to be rated higher than pensions by younger workers. Lower seniority status concomitant with cyclical and seasonal output fluctuations incur greater layoff pos- sibilities to the young. Yet, for workers under age 30, pensions almost equalled the combined ratings of both for steelworkers, and almost doubled their combined ratings in the non-steelworker sample. One possibility for the high rating of pensions sug- gested by Wagner and Bakerman, is that a pension is one fringe benefit that the worker hopes to collect. All other fringe benefits covered in this study, with the ex- ception of vacations, are benefits which usually occur because of a calamity. It is possible that man does not predict misfortune for himself; therefore, benefits such as life and health insurance, unemployment compensation, etc, might have a negative appeal lacking the strength of blissful Florida retirement. Also, with increasing life expectancy coupled with a trend towards earlier retire- ment, the probability of having a longer retirement life is rising. 33 Also, pension credit compensation, if vested, is a form of saving. Unionized workers in the primary metals industry are at income levels well above the mean average income of workers in the nation. Therefore, if the pro- pensity to save is higher among workers in the higher income brackets, then one might expect such workers to more readily accept income increments in the form of pen- sion credits than lower income workers. Furthermore, pension credit compensation is not considered income by the Internal Revenue Bureau and, therefore, is not taxable as is the case with wages. Although the income from pen- sions is taxable upon retirement, workers stand to gain by deferring such tax payments. In summary, although the Wagner and Bakerman study did not specifically attempt to test the pension-mobility hypothesis, it did offer some evidence to validate the hypothesis. Workers preferred increases in pension bene— fits over increases in regular wages. Evidence contrary to the Wagner and Bakerman find- ings regarding the strength of pensions was found by Herbert Parnes.5 From February through July of 1958, Parnes conducted interviews with a sample of workers em- ployed at two metal-fabricating firms in Columbus, Ohio. The output, working conditions, wages, characteristics of 5Parnes, "Workers Attitudes to Job Changing. . .," pp. A5-80, 182-189. 3A employees, and other conditions were similar in both establishments, the fundamental difference being that one provided its employees with a pension plan, while the other did not.6 Among several queries put to the workers, one di— rectly pertained to the pension-mobility hypothesis. Respondents were asked whether they would accept an iden- tical position elsewhere in Columbus at 30 cents an hour above the present wage. Unqualified refusals came from 59 per cent of the 97 workers interviewed from the non- pension plan firm and from 60 per cent of the 93 workers in the firm with a pension plan. When asked why they re- fused, none of the respondents mentioned the potential loss of pension rights. With regard to why they might return to work at their present firm, after being laid off and having an identical job elsewhere in Columbus, over 50 per cent of the workers employed by the firm hav— ing the pension plan said that they would return. None of the respondents mentioned the pension plan as a reason for returning. Parnes concluded, mainly on the strength. of the responses to these two questions, that pensions 6Ibid., p. 185. Employees, upon retirement at age 65, received a monthly benefit equal to (a) $.50 times-the number of years of service until age A0, plus (b) $1.00 times the number of years of service from age A0 to 65. The foregoing benefit formula was doubled just prior to Parnes' study. Pension credits vested after 15 years of service. 35 were of little importance in attaching production workers to their jobs.7 It is necessary to make but a single comment concern- ing the merit of Parnes' pension-mobility study. If pen— sion credit losses are to be an inhibiting factor to the mobility of labor, their value must be relatively import- ant to the employee. The pension plan in Parnes' study was initiated by the firm in January, 1956, and the bene- fit formula was liberalized in early 1958, resulting in a 100 per cent benefit increase. All respondents were in the 35 to 50 age bracket. The maximum monthly retirement loss at age 65 that any respondent could incur if he quit his job at the time of the interview was $3.00 per month.8 It seems inconceivable that a $3.00 per month loss at age 65 would impede any worker from leaving a job, let alone workers 50 years of age and under. The most important condition for an accurate test of the pension—mobility hypothesis was absent in the Parnes' test, i.e., non-vested credits must be signifi- cantly high enough to be of importance to the individual employee. Parnes' study, regarding the mobility inhibit— ing effect of pension plans, borders on sophistry. 7Ibid., p. 77. 8n. 6. Assuming age A0 when the plan began in Janu- ary, 1956, the $3.00 per month benefit was computed as follows: $1.00 per month times two years of service plus $2.00 per month times one-half year's service. 36 In Oregon, in 1962, an opinion survey of faculty members of State Institutions of Higher Education was conducted by Mark Greene.9 There were 1,850 replies to a mail questionnaire covering approximately 2,500 faculty members. Respondents were covered by the Oregon Public Employees Retirement System. The plan is contributory, with the state matching the employee's six per cent con- tribution. The employer's contribution vests after ten years of service.10 Although Greene did not specify the exact phrasing, respondents were asked questions concerning the import- ance of delayed vesting with respect to whether they would accept an offer from another institution. Table A summarizes the replies to such questions. The data suggest that length of service had little effect on the strength of non-vested credits in inhibiting labor mobility. Greene, however, did not specify what questions were asked but merely stated that employees were asked questions concerning the importance of vesting to them. If employees' responses were based upon their pre- sent situation, then it is interesting to find that fac- ulty having ten or more years of service would consider non-vesting a problem since the Oregon plan vests at the 9Mark Greene, "Fringe Benefits or Salary," Journal of Marketing (October, 1963), pp. 63-68. 10Ibid., p. 6A. 37 mo m0 mcfipmo> wcfizmm ucoo pom .00 .Q =.0smamm so whammcmm mmcfinms .ocoomo Song Cosme "condom 0000 0.00 0.00 0.00 m.w 0.00 Hopos 000 0.00 0.00 0.00 0.0 0.00 meso0 00 noeo 000 0.00 0.00 0.00 0.00 0.00 memo» 00-00 000 0.00 0.00 0.00 0.0 0.00 00000 00-00 mmm 0.00 0.00 0.00 0.0 0.00 munch m|0 000 0.00 0.00 0.00 0.0 0.0 00000 0-0 00 0.0 0.00 0.00 0.0 0.00 steam no: oo0>hmm Hmpoe mocmppoaefi oosmpsogefi conspsogefi mocmpsomefi Loamcm oanMom oz ofipufiq 00000002 names 02 npwsoq .nommo oofimuso cm udooom o» coamfiooo w :0 mcfipmm> commaop mo oocmpnoaeHnl.: mqm¢e 38 end of 10 years of service. If they were asked to assume that their pension credits were non—vested, then it is equally interesting to find such little importance placed upon non-vested credits, by many long service employees. Thus, it seems that respondents differed in their inter- pretation of Greene's questions. Some respondents answered with respect to their present employment within the Oregon plan, whereas others assumed they were not to consider their present vested or non-vested position. Thus, Greene's findings seem irrelevant to the general problem raised in this chapter. In a 196A survey, Harold Rubin studied the impact of pension plans on the mobility of public employees in pro- fessional jhb classifications.ll Unlike the studies which solicited a sample of employees' opinions, Rubin restricted his inquiry to new personnel.l2 By means of a mail questionnaire, employers were re- quested to provide data concerning (1) the number of appli- cants who declined offers, (2) the number who specifically mentioned pensions as a factor in their declination, and (3) the number of salary increases offered to applicants llHarold Rubin, Pensions and Employee Mobility in the Public Service. l2Ibid., pp. vii-viii. Public agencies surveyed in- cluded health, welfare, education, employment security, and agriculture, in the five largest cities and five larg- est states in the nation. Also, a selectibn of public col- leges and universities in California, New York, Pennsyl- vania, Illinois, and Ohio was included. 39 in an effort to offset pension credit losses.13 Thus, Rubin tested a combination of two impediments: (1) the restraining influence that non-vested plans impart in the hiring of personnel; and (2) the restraining influence that non-vested plans have on personnel considering termi- nation. Relatively few employers reported pensions as a factor in job offer refusals, and still fewer reported that salary increases had been used to offset employee pension losses.lLl Two factors, however, cast doubt as to the accuracy of the respondents' reports. First, the low incidence of the pension factor as a condition of employment may have been due to an absence of accurate record keeping by employing agencies. For ex- ample, 8 out of 51 institutions of higher education in the State University of New York System accounted for all of the job declinations where pensions were a factor. All New York State Colleges and Universities are covered by the same retirement plan. It seems improbable that A3 of the 51 responding institutions would not have any pension 15 declination cases. Furthermore, the experience of one University in the New York State System was revealing. 13Ibid., p. 9A. l“Ibid., pp. 21—62. In general, there was no sig- nificant difference between the educational and non— educational responses. lSlEiQ-s p- 25. The 51 institutions of higher edu— cation in the New York State System responding to the A0 The State University at Buffalo had been the pri- vate University of Buffalo until 1962, at which time it joined the New York State System. When privately con— trolled, the university had a TIAA-CREF retirement plan which was discontinued at the time of entrance into the New York State System. After becoming a part of the New York State System, the State University at Buffalo con- tinued its policy of recruiting faculty from private col- leges and universities where TIAA-CREF is the predominant plan. Three-fourths of the reported declinations where pensions were a factor were accounted for at the State University of Buffalo. Approximately 50 per cent of the rejected offers at Buffalo were wholly or partly attri- buted to the pension factor.l6 Since TIAA-CREF plans are fully vested, it cannot be assumed that applicants re- fused positions at Buffalo because of the loss of pension credits due to separation from their present schools. Rather, it seems that two factors must have been responsi— ble: (1) the inability to continue contributing to TIAA- CREF; and (2) the 15-year vesting period in the New York Retirement System. The second reason that casts doubt on the statis- tically reported overall irrelevance of the pension factor questionnaire reported 600 job refusals and approximately 1 of 8 of the 600 refusals were partially or wholly re- lated to pensions. 16Ibid., p. 25. Al as an inhibiting force in the hiring of employees is the respondent's reply to a subjective question. The re- spondents were invited to give their opinions on the desirability of liberalizing vesting and transfer provi- sions in their retirement plan. More than 90 per cent of the respondents favored vesting and/or transfer provision liberalization. However, almost 75 per cent believed that liberalization would not affect employee turnover.17 It seems that many of the respondents believed that vest- ing and transfer provision liberalization would be advan- tageous in hiring sought after personnel, but that it would not noticeably increase the rate of voluntary quits from their respective institutions. If such a high per- centage of respondents wanted vesting and/or transfer provisions to be liberalized at their institutions in order to ease the hiring problem, then why did so few re— port that pensions were a factor in job offer refusals? The respondents' reports contain a measure of self— contradiction. Rubin's findings concerning the experience of the University of Buffalo are enlightening. Pension plans may give rise to a fractionization of the labor market, especially of the professional labor market. Profes- sional employees, if aware of the merits of fully vested plans and the relatively fast accrual of returns on 17Ibid., p. 68. A2 contributions in such plans as TIAA—CREF, may limit their marketability to institutions having such plans. Quit-Rates and Non—vested Funds Melvin Lurie tested the pension-mobility hypothesis by comparing the voluntary quit—rates of college and uni- versity faculty employed at non—vested schools YE- the voluntary quit-rates of faculty from vested schools.l8 Lurie obtained pension and resignation data for faculty members in 10 departments at 339 colleges and 98 universi- ties for the academic year 1959. By arbitrarily defining plans that did not fully vest within five years as non- vested, the sample contained A7 per cent vested schools and 53 per cent non—vested. Because of a variety of fac- tors, other than loss of pension credits which could in- fluence faculty resignations, the sample was standardized to attempt to isolate the pension variable.19 18Lurie,"The Effect of Non-Vested Pensions on Mobility . . .," pp. 22A—237. 19Lurie's "standardization" procedure was as follows: Within each university, factors such as geographic loca- tion, administrative control, research facilities, etc., would have approximately the same mobility affect on all faculty. However, faculty within an institution would have varying pension credit values. The greater the value of non-vested credits, the greater the impediment to quit- ting. Lurie did not have data on the value of pension credits, however, he assumed a service-pension credit value relationship, i.e., the longer the service, the greater the value of pension credits. Having no data on faculty length of service, he substituted rank for length of service. Thus, a double substitution was made, years of service for value of pension credits and rank for years of service. The higher the rank, the stronger the A3 The statistical results of Lurie's survey for all A37 institutions are summarized in Table 5. TABLE 5.—-Voluntary separation rates, according to length of covered service, in vested and non-vested institutions of higher education, 1959. Ratio of Relative Covered Vested to Service Vested Non-Vested Non-vested (sensitivity ratio) Professor/Associate Professor .311 .526 .591 Professor/Assistant Professor .272 .273 .996 Professor/ Instructor .135 .156 .865 Source: Melvin Lurie, "The Effect of Non-Vested Pensions on Mobility . . .," pp. 22A—237. If the value of non-vested credits inhibit faculty from leaving non-vested plan institutions, then the ratio of quit-rates in non-vested institutions should be lower than the quit-rate ratios in vested institutions. The data in Table 5 depicts the opposite. All the sensitivity ratios should be greater than unity if the pension— mobility hypothesis is to be validated. As the data indi- cate, all ratios are less than unity; therefore, the impediment to mobility within a non-vested institution. By dividing the quit-rate of professors by the quit-rate of each lower rank within an institution, Lurie obtained comparison ratios for each institution. AA results seemingly invalidate the pension—mobility hypo- thesis. However, Lurie arbitrarily defined all plans that did not vest within five years as non-vested. Many insti- tutions have plans that at the time of Lurie's survey, vest after 10 or 15 years of covered service.20 In such schools, a far higher percentage of full professors than those of lower rank would have attained a fully vested position. Therefore, if many such institutions were in— cluded in Lurie's sample, the results depicted in Table 5 may attest to the validity of the hypothesis rather than invalidate it. Certainly the low professor/associate pro- fessor sensitivity ratio, .591, attests to the validity of this possibility. It is logical to assume that a higher percentage of associates employed in schools that defer vesting are more proximate to obtaining full vesting than any other rank. Therefore, as a group, associates would be more tightly "locked in" than any of the other three ranks.21 Lurie disaggregated his vested and non-vested insti— tution samples into college—university sub-samples and further sub-divided by type of administrative control. The results of the sub-division are depicted in Table 6. 2OSee, William C. Greenough and Francis P. King, Retirement and Insurance Plans in American Colleges (New York: Columbia University Press, 1959). 21See the analysis on deferred vesting in Chapter II. A5 TABLE 6.--Facu1ty sensitivity ratios, according to length of covered service, in colleges and universities by type of control, 1959. Relative Covered Colleges Universities Service Private Public Private Public Professor/Associate 1.5A7 * A.688 .65A Professor/Assistant .986 .010 1.975 1.6A8 Professor/Instructor .609 .002 2.279 1.01A *No voluntary separations at rank of associate professor among publicly controlled vested colleges. Source: Lurie, "The Effect of Non-Vested Pensions on Mobility . . .," p. 230. Note that the sensitivity ratios for universities, both public and private, are all greater than unity, with the exception of the professor/associate professor ratio for publicly controlled universities. This single excep- tion again, given the number of public universities that vest after 10 to 15 years of service, may be a validation of the hypothesis rather than an invalidation. Also, there seems to be some inconsistency in Lurie's data, particularly the professor/associate professor sensitivity ratios. All of these ratios are larger than the aggregate ratio reported in Table 5.22 The data in Table 5 are the aggregated information depicted in Table 6. How can each 22The professor/associate professor sensitivity ratio for publicly controlled colleges would be approach- ing infinity. A6 of the sub-divided samples have sensitivity ratios greater than the composite sample? This does not seem mathemati- cally possible; the apparent contradiction, coupled with the inclusion of plans that vest after five years in the non-vested sample, seriously weakens the significance of Lurie's results. ' In 1958, a study often quoted as shedding light on the pension—mobility hypothesis, was conducted by Arthur Ross.23 The purpose of Ross's study was to ascertain whether the growth of seniority systems, health and wel- fare plans, and pension plans, had the effect of tying workers to their present positions. Ross compared yearly quit-rates in manufacturing em- ployment for the years 1910-1956. After adjusting for changes in the level of unemployment, a moderate reduc- tion in the quit—rate was discernable in the 1950's. Ross ascribed the reduction in the manufacturing quit-rate to the following: (1) improved working conditions, (2) the aging of the labor force, (3) the stability of manu- facturing employment, and (A) the effect of seniority. rules in protecting the worker against replacement by a new worker. Ross did not directly challenge the pension- mobility hypothesis, rather he implied its invalidation 23Ross, "Do We Have a New Industrial Feudalism," pp. 903,920. A7 by omitting pension plans as a cause in the declination of manufacturing employment quit-rates. Although the factors suggested by Ross as influenc- ing the decline in quit-rates have merit, they cannot be considered all-inclusive. By merely omitting the pension variable in his subjective analysis, Ross did not negate its influence.2Ll Although Ross disclaimed the mobility inhibiting influence of non—vested pension plans, his procedure did not isolate the pension variable.and, therefore, cannot be considered a test of the impact of pensions on labor mobility. Concluding Remarks The studies reviewed in this chapter have, on the whole, cast doubt on the validity of the pension-mobility hypothesis. However, the accuracy of findings are subject to rather damaging criticism. Wagner and Bakerman's sur- vey, although giving some insight into employees' aware- ness of pension benefits, did not test the mobility re- straining influence of pension benefits. Parnes' sample selection erased the possibility of his obtaining mean— ingful results. The responses to Rubin's queries were logically contradictory and Greene's results suggest re- spondents misinterpreted his questionnaire. Lurie's 2”Two other important factors omitted by Ross are increasing industry concentration and government economic policies. A8 inclusion of plans that vested after five years in his non-vested sample was akin to testing the braking power of a car by concurrently applying pressure on the acceler- ator as well as the brakes. Ross's test cannot be in- ferred as a test of the pension-mobility hypothesis since his procedure did not isolate the pension variable. Thus, it seems that little progress has been made concerning the empirical validation or invalidation of the hypothesis. CHAPTER IV EMPIRICAL TEST OF THE PENSION— MOBILITY HYPOTHESIS Two general approaches typify the literature deal- ing with tests of the pension—mobility hypothesis. First, through questionnaires and personal interviews, investi- gators have endeavored to discover how cognizant workers are of their pension plans, and how important plan bene— fits are relative to other benefits in keeping workers in their jobs. Although this investigative technique is pro— cedurely sound in isolating the pension variable, it does not directly measure the mobility inhibiting strength of non-vested pension credits. Rather, the investigator must translate workers' opinions of the value of pensions into mobility-impediment strength of non-vested plans. The second technique involves comparison of quit- rates from non—vested employment with quit—rates where non-vested plans do not exist. This procedure does not suffer the opinion—method problem of transferring opinion to actual mobility-inhibiting strength of non-vested plans. It does pose greater difficulty in isolating the pension A9 50 variable, but by careful selection of samples, the influ— ence of factors (other than non-vested pension credits) that might affect mobility can be mitigated. It is this procedure (specifically, comparison of voluntary separa- tion rates of employees from non-vested employment with voluntary separation rates from vested employment) that will be used to test the validity of the hypothesis that the form of the pension plan affects mobility of univer- sity faculty. Sample Selection There are three requisites for a quit-rate test of the hypothesis: (1) a non-vested test sample, (2) a control sample, and (3) quit-rate data for both samples. Specific requirements that directly influence the selec- tion of samples were as follows. 1. The non-vested plan selected must not be an atypical non—vested plan if the results of the test are to have general projective significance. Although pen- sion plans are widely divergent with regard to participa- tion, benefit requirements, and type and amount of bene- fits, the provision that is universal to all is the retirement benefit provision. The loss of the present value of retirement benefits is the major compensation loss incurred by the non-vested employee upon separation. Since the present value of non-vested credits is directly 51 related to benefits received at retirement, the investi- gator can destroy the projective applicability of his test results by selecting a non—vested plan sample yield- ing relatively low or high retirement benefits. In a recent Bureau of Labor Statistics survey of pension plans, retirement benefits among plans ranged from less than $10 per month to over $AOO per month for participants having identical service and earnings histories.l Results ob- tained from a test utilizing data from a non-vested sample at either of these extremes would not be relevant to the economy. Therefore, in order to project test results, a primary criterion directing the selection of the non- vested sample was avoidance of plans paying atypical re- tirement benefits. 2. Since length of service is directly related to benefits in almost all plans, a second criterion directing the choice of the non—vested sample must be the life of the plan. Non-vested plans that meet the first criterion will not meet this second criterion if plan participants have not built up non-vested credit values to the point where their loss might impede voluntary separations.2 Therefore, the second criterion that directed the choice 1U. S. Department of Labor, Bureau of Labor Statis- tics, Private Pension Plan Benefits, Bulletin Number 1A85 (June, 1966), Table 10, p. 32. 2This is a major criticism of both Parnes' and Ross's studies, discussed in the last chapter. S2 of the non-vested sample in this study was avoidance of plans with a limited history. 3. Factors recognized as affecting worker mobility may be broadly categorized under labor market conditions and personal characteristics of workers. More specifi— cally, workers will tend to move into occupations, indus- tries, regions, etc., that yield higher compensation, increased security, and better chance for advancement. Personal characteristics such as age, race, sex, marital status, education, and income, have recognizable influence on labor mobility. To minimize the influence of such variables and, therefore, to isolate the effect of non- vested pension credits on mobility, the vested sample must be as homogeneous as possible to the non—vested sample in all particulars affecting mobility. A. Both samples must be large enough to insure statistically relevant results. 5. Finally, accurate voluntary separation data for both samples must be obtainable. Non-vested Sample With the aforementioned criteria as a guide, the non-vested sample selected for this study was comprised of faculty members, assistant professor and above, em- ployed in ten departments at each of the six public 53 universities in the State of Illinois.3 The ten depart- ments were chosen on the basis of their common occurrence in American universities: Chemistry, Economics, English, History, Mathematics, Philosophy, Physics, Political Science, Psychology, and Sociology. All faculty members included in the non—vested sample were participants in the State Universities Retirement System of the State of Illinois. Faculty participation in the Universities Retirement System of Illinois is voluntary for the first year of em- ployment and mandatory beginning the second year. The plan provides benefits for retirement, survivors, and dis- ability. Retirement and survivors' benefits, for the period in which the test was run, vest after ten years of participation.Ll Monthly retirement benefits are defined by the fol- lowing formula: 1—2/3 per cent times number of years of service, times final rate of earnings. "Final rate of earnings" is defined as the average yearly earnings (in- cluding summer session and all extras) during the five 3The six universities included are: (l) The Univer- sity of Illinois (Urbana campus), (2) Southern Illinois University (Carbondale campus), (3) Northern Illinois Uni- versity, (A) Western Illinois University, (5) Eastern Illinois University, and (6) Illinois State University. ”The period for the test included the academic years l96A-65, 1965-66, and 1966-67. Beginning with the academic year 1967-68, the vesting period was reduced to five years. 5A consecutive fiscal years in which earnings were the high- est. The plan is contributive, with the employee contri- buting six per cent of his monthly earnings as his share in the cost of retirement benefits. The participant can begin receiving full retirement benefits at age sixty unless he receives compensation from an employer covered by the State Universities Retirement System of Illinois, the State Employees' Retirement System of Illinois, or the Teachers' Retirement System of Illinois. Compensa- tion received by the participant, during any month, in excess of the retirement annuity for services performed for any employer covered by these retirement systems, is subtracted from the State's share of the monthly annuity payment. The participant, however, receives that portion of the annuity provided by his contribution and interest. Thus, a participant, having obtained a vested position, can receive full monthly benefits at age sixty, while working for employers not participating in any of the three Illinois pension systems. This feature cf the Uni- versities Retirement Plan would seem to encourage volun— tary separations upon obtaining vesting privileges. Annuity payments, and a lump-sum of $1,000, are received by survivors upon the death of a participant. No single formula defines annuity payments, however, Table 7 illustrates more common survivor benefit schedules. 55 TABLE 7.—-I11ustration of monthly survivors benefits. ‘ Monthly Final Widow, Dependent Widow and Widow and 2 Rate Of Husband, OP 1 Child or more Earnin s Dependent Parent (60 per Children g (30 per cent) cent) (80 per cent) $100 $ 30 $ 60 $ 80 200 60 120 160 300 90 180 200 A00 120 2A0 250* 500 150 250* 250* 600 180 250* 250* 666.67 200* 250* 250* *Maximum Source: State Universities Retirement System of Illinois, Handbook of Information, State Universities Retirement System (Urbana, Illinois: 1965), p. 31. Survivor's annuity insurance is mandatory for all retirement plan participants. Costs of the survivor plan are shared by participants and the State of Illinois, with the participants contributing one per cent of gross earnings per month with a ceiling of $80 per year. The State of Illinois assumes all remaining costs. The Actuary of the Illinois State Universities Re- tirement System has estimated the employer's share in the cost of the total plan at ten per cent of total partici- pant payroll.5 5State Universities Retirement System of Illinois, Handbook of Information, State Universities Retirement System (Urbana: 1965), p. A8. 56 Non-vested credit values.--A1though survivors' an- nuity benefits in the Illinois plan are vested after ten years of service, the factors that might influence parti— cipants'non-vested credit values are too numerous to permit objective appraisals of their possible mobility inhibiting strength.6 It is sufficient to say that the value of non- ‘vested survivor annuity credits almost certainly has some ianediment effect upon mobility, however, it probably has fuar less effect than non-vested retirement credits.7 It is the value of non-vested credits that offers tile possibility of significant curtailment in the rate of v<>luntary separations. Provisions in the retirement plan affecting non-vested retirement credit value are: (l) ruanevested credits vest after ten years of service, (2) enuployees contribute six per cent of gross earnings each nuxnth into the retirement plan, (3) terminating partici- paJTts receive a refund consisting of their contribution P1413 3-1/2 per cent interest, compounded annually, and —; 6The value of survivors' annuities depends upon: (1) the final rate of earnings of the participant, (2) thfi! number of dependents, and (3) the ages of dependents. Adi! to these variables those which influence the proba— bility of death during the period in which the partici— Paxrt has many dependents Kg, periods in which there might be few or none. 7This statement assumes that the employer's share CW7 survivor payments and retirement payments are propor- tional to the employee's contributions from earnings of one per cent or less and six per cent respectively, for these benefits. 57 (A) vested participants receive a yearly retirement in- come at age 60 equal to l-2/3 per cent times years of service times final rate of earnings. Since retirement benefits vary with service and earnings, non-vested credit values will vary with changes in each of these parameters. Furthermore, as illustrated in Chapter II, the age of plan participants and the interest rate used to compute present values of pension credits influence the value of non- vested credits. To calculate the non-vested credit values E for all possible value combinations of the parameters would be an insurmountable task. An example computation, however, will reveal the implications of changes in the parameters. Assume the following: 1. A male participant's initial salary is $8,000 per year. 2. His salary increases at the rate of 6 per cent per year. 3. The interest rate for all calculations is 3-1/2 per cent per year. Utilizing the benefit formula and the assumptions, the lump-sum value of the participant's annuity at age 60, upon completion of ten years of service, is $27,258.8 By 8The participant at age 60 will receive an annuity of $167.67 per month. The selling price, common to large insurance companies, of such an annuity is $27,258. See, Dan McGill, Fundamentals of Private Pensions (2d ed.; Homewood, Illinois: Richard D. Irwin, Inc., 196A), p. 90. 58 discounting the lump-sum value of the employer's share at age 60 to various participant age and service positions, one can compute the yearly remunerative value of non— vested credits to age 60. The values illustrated in Table 8 represent the yearly remuneration increase (to retire- ment at age 60) required to offset the non-vested credit loss if the participant leaves the plan. TABLE 8.——Non—vested credit value per year to age sixty. Vesting at 35 Vesting at A5 Vesting at 55 Age Service Value Age Service Value Age Service Value 26 1 $158 36 1 $A10 A6 1 $1,051 27 2 166 37 2 A36 A7 2 1,153 28 3 17A 38 3 A65 A8 3 1,273 29 A 18A 39 A A96 A9 A l,A1A 3O 5 19A A0 5 531 50 5 1,58A 31 6 20A A1 6 570 51 6 1,792 32 7 216 A2 7 613 52 7 2,053 33 8 229 A3 8 662 53 8 2,389 3A 9 2A2 AA 9 716 5A 9 2,837 35 10 257 A5 10 779 55 10 3,A66 An example will clarify the method used to obtain the values in Table 8. Assume the participant joined the plan on his 35th birthday. He would have contributed $A80 during his first year of participation, $509 during his second year, etc. By taking each year's contribution separately to retirement at 3-1/2 per cent compound in- terest, the value of his contributions at retirement is $12,235. Therefore, the employer's share is equal to 59 $27,258 - $12,235, or $15,023, at age 60. At 3—1/2 per cent return on the year-end balance, the participant at age AA would have to receive $716 per year for 16 years to obtain $15,023 at age 60.9 Observe that the plan favors the participant who begins his employment at a later age. Although contribut— ing the same amount of money, the newly vested participant i separating at age 35 receives the same benefit at age 60 as the participant who separates upon reaching vesting at 3 % age 55. This results from the fact that the retirement benefit formula does not credit interest on the partici- pant's contribution if left in the plan upon separation. Note also that the interest rate used for all compu- tations was 3-1/2 per cent, the rate paid by the state on employee's contributions refunded upon separation. If the employee believes he can obtain a higher rate on his contributions upon termination, it_might profit him to take the refund even if he has attained a vested position. For example, the value of a participant's contributions, including interest at 3-1/2 per cent, is $7,302 at vesting. If he separates at age 35, he has the option of leaving the $7,302 in the fund and collecting an annuity valued at 9The procedure used, although differing from that presented in Chapter II, yields the same result. By tak- ing the additional step of calculating the present values of non-vested credits for any age-service combination, the yearly remunerative values to retirement can be com- puted in accordance with condition (7) in Chapter II, and will equal those depicted in Table 8. 60 $27,258 at age 60, or he can take the refund. The in— terest rate per year on the $7,302 refund that would have to be earned for 25 years in order to equal $27,258 at age 60 is approximately 5-3/8 per cent. If a participant vests at age A5 and separates, the rate of return on the refund would have to be approximately 9—1/A per cent; and at age 55, about 29-3/8 per cent. When one considers the added incentive of having access to the contributions, non—vested retirement credit losses should not be a sig- nificant hindrance to mobility to younger participants. Older non-vested participants, however, might hesitate to consider even casually a voluntary separation in view of the retirement losses that might be incurred. Although outside offers lower in yearly differential remunerative value than those values depicted in Table 8 should be rejected by participants, acceptance of outside offers with differentials higher than the values in the table might not be the best choice of the participant. Discussion, to this point, has involved only two possible choices: (1) the non—vested participant could remain in his present position until retirement, or (2) the non- vested participant could accept a more remunerative posi- tion although losing the non-vested credits. A third choice, however, is available to participants. They might reject the superior outside offer in the belief that another offer will arise after a vested position has been 61 attained. This third possibility was discussed in Chapter II, and it culminated in formulation of condition (8) in that chapter. Illustration of this third choice makes it necessary to add the assumption of a specific offer to all previous assumptions used in computing the values in Table 8. Table 9 depicts yearly remuneration values from the vesting date to retirement necessary to compensate the participant upon rejection of an offer, during the non- vested period, that is $1,000 per year higher to retire- ment than would be received in his present position. TABLE 9.--Higher offer necessary at vesting to equal $1,000 higher offer during various non-vested periods. Vesting at 35 Vesting at A5 Age Service Value Age Service Value 26 1 $1,372 36 1 $1,121 27 2 1,292 37 2 1,009 28 3 1,215 38 3 896 29 A 1,1A0 39 A 790 30 5 1,068 A0 5 687 31 6 999 A1 6 587 32 7 931 A2 7 A91 33 8 866 A3 8 398 3A 9 80A AA 9 308 The values in Table 9 were calculated as follows. Assume age A3 and completion of the eighth year of service. The value of non—vested credits at the time of vesting, age A5, is $8,972. The differential that the participant must 62 receive upon completion of the tenth year equaling the $1,000 per year offer over his present renumeration at the end of the eighth year is $398.10 Therefore, if the participant believes that he can obtain an offer at vest- ing that will be higher than $398 per year each year for 15 years than his expected income in his present position he should not accept the outside offer at the end of his eighth year of participation. Note that an increase in the interest rate or an increase in the offer differential prior to vesting would increase the values in Table 9. For example, a differen- tial increase of $2,000 per year rather than $1,000 per year would raise the renumeration differential required at vesting for the A3 year-old participant from $398 to $1,57A per year. Because of the many parameters that influence non— vested credit value, it would be meaningless to specify a precise age-service combination where such values become a significant impediment to mobility. The results of the analysis, however, accent the importance of the age at loUtilizing condition (8), we have: /' \ 2 $1,000 (1 + .035) - 1 $ $8,972 - .035 R = 1,000 - l — (l + .035)‘15 .035 K / R = $398 63 which participants begin their service.11 It is doubtful if losses would be a significant deterrent to non-vested participants that joined the plan before the age of thirty. Those that become employed in their thirties have some concern when considering outside offers, while those employees who become participants in their forties might 0w . well feel "locked in." Vested Sample The vested sample included faculty members, assist- g ' ant professor rank and above, at Iowa, Iowa State, Michi- gan State, and Purdue Universities employed in the ten departments included in the non-vested sample. Each university in the vested sample provides pen- sion coverage for faculty members through the Teachers Insurance and Annuity Association and College Retirement Equities Fund (henceforth referred to as TIAA-CREF). Benefits include retirement and survivor's income and are determined by the employer's and participant's 11For example, the yearly remuneration values of non-vested credits to retirement for participants com- pleting the first year of service and starting at ages A5 through A9 are: Age A5, $1,051 per year. Age A6, $1,172 per year. Age A7, $1,312 per year. Age A8, $1,A78 per year. Age A9, $1,679 per year. One might accurately describe the Illinois plan as a "windfall" for the middle-aged entrant and a "penalty" for the young entrant. 6A contributions to the participant's account with TIAA-CREF, coupled with earnings from the total TIAA—CREF investment portfolio. Of singular importance for this study is the fact that employer's contributions are deposited in the participant's TIAA—CREF account and, thus, are immediately and fully vested to the participant. Therefore, the fac— ulty members in the vested sample do not suffer the loss a of employer's contributions upon employment termination. Comparison of the Samples It is clear that there are factors, other than the extent of pension vesting, that affect the quit—rates of faculty. To attempt to identify and measure the absolute impact on mobility of such factors is not necessary for this study. It is necessary, however, to consider those factors that might give rise to a difference in the quit- rates of the samples. Two such factors are the prestige of the institutions and the rate of faculty compensation. It would be generally recognized by faculties through- out the nation that the institutions from which the vested sample was drawn are, as a group, more prestigious than those institutions from which the non-vested sample was selected. Assuming faculty tend to gravitate toward the more prestigious schools, this factor might have a dampen- ing impact on the aggregate quit-rates of the vested sample relative to the non—vested. 65 The second factor, compensation differences between the samples, may have a substantial impact on the relative aggregate quit-rates. The average compensation at the. sample non-vested universities and the vested universi- ties, for the test period, is shown in Table 10. TABLE lO.--Average compensation, full—time faculty. ’ Universities l96A-65 1965-66 1966-67 II Vested $11,361 $12,A82 $13,2A5 Non-vested* 9,585 10,196 10,652 Differential 1,776 2,286 2,593 *Pension plan remuneration values are not included in the non-vested sample compensations. Source: American Association of University Pro- fessors, AAUP Bulletins, Summer Issues (Washington: 1965, 1966, 1967). Although the addition of the average compensation value of the pension plan would reduce the differential exhibited in Table 10, the average compensation of faculty from the vested schools would still be far above that re- ceived by faculty at schools from which the non-vested 12 sample was drawn. Given that labor seeks higher remun- eration for services rendered, we might expect that faculty 12The value of non—vested credits to faculty partici— pating in the Universities Retirement System of Illinois will be discussed later in this chapter. 66 in the non-vested sample would tend to voluntarily sepa- rate more readily than faculty in the vested sample. Note, however, that all participants in the non— vested sample will be equally affected by the difference in prestige and compensation variables. Therefore, the effect of these variables would fall on aggregage compari— sons of quit-rates, and should not affect comparisons that allow for the strength of non-vested credits within the non-vested sample. Additional comments on the affect of the compensation variable will be found below. The data collected for the test were as follows: 1. The names of all faculty members, assistant professor rank and above, in each of the ten departments at all ten universities. 2. The date at which the faculty member was first employed with the university. 3. The last year employed for those faculty mem- bers who voluntarily terminated. The collection procedure involved both mail and per- sonal interviews with administrators, faculty, and staff at each university. Information was collected covering 98 departments (59 in the sample of vested plans and 39 in the non—vested).l3 13Illinois State University did not have-a Philosophy Department during the test period, and the data from the Physics Department at Michigan State had not arrived at the time of this writing. Therefore, the returns from 98 departments comprise the total test data. I ‘5 67 If non-vested pension credits inhibit the mobility of workers, the voluntary separation rates of faculty in the vested sample should exceed the voluntary separation rates of faculty in the non-vested sample. Table 11 summarizes the quit-rates for both samples by academic year. Assuming homogeneity of all factors that influence mobility other than the pension plans, the non-vested sample quit-rates should be less than the vested sample quit—rates. The data depicts the reverse. Note, however, that total separations, by year, were included in the computation of separation rates for both samples. The Illinois plan vests after ten years of service, therefore, the quit rate data for the non-vested sample includes faculty who terminated after attaining vested status.lu To remove the influence of vested separa- tions on the non-vested separation rates, voluntary separa- tions in both samples were restricted to the non-vested period. The separation rates for the non—vested period are depicted in Table 12. Although the separation rate for the academic year 1966-67 is smaller for the non-vested sample relative to the vested, the separation rate total for the three year 1”The inclusion of vested faculty as non-vested in the computation of quit-rates was a basic criticism of Lurie's procedure as was noted in the last chapter. 68 TABLE 11.--Voluntary separation rates, non-vested sample vs. vested sample. Non-vested Plan Vested Plan Year Number of Number in Separation Number of Number in Separation Separations Sample Rate Separations Sample Rate 196A-65 62 1,000 6.20% 52 1,082 A.80% 1965-66 59 1,108 5.32% 52 1,169 A.A5% 1966-67 93 1,222 7.61% 8A 1,251 6.71% Total 21A 3,330‘ 6.A3$ 188 3.502 5.37% TABLE 12.—-Vo1untary separation rates during the non-vested period. Non—vested Vested Year Number of Number in Separation Number of Number in Separation Separations Sample Rate Separations Sample Rate 196A-65 51 6A9 7.85% AA 625 7.0A% 1965—66 56 7A5 7.52% AA 683 6.AA% 1966-67 75 836 8.97% 67 721 9.29% Total 182 2,230 8.16% 155 2,029 7.6A% 69 period is larger for the non-vested sample. Though the size of the differential is less than that depicted in Table 11, voluntary separation rates were still higher in the non-vested sample. It should be understood, however, that the data in Table 12 conceals the strength of the basic inhibiting ingredient of the non-vested plan, the value of the non—vested credits. The value of non-vested credits increases as the participant approaches fully vested status, therefore, voluntary separation rates should diminish faster in the non-vested sample as com- pared to the vested sample. Table 13 depicts the volun- tary separation rates by year of covered service and Figure 1 illustrates the quit rate trends. TABLE l3.--Voluntary separation rates by year of partici- pation. Non—vested Vested End-of- year Number of Total Per Number of Total Per Separations Faculty Cent Separations Faculty Cent 1 A6 51A 8.95 21 A17 5.0A 2 38 376 10.11 28 366 7.65 3 25 268 9.33 36 286 12.59 A 21 253 8.30 26 215 12.09 5 27 227 11.89 7 175 A.00 6 6 191 3.1A 12 1A2 8.A5 7 9 153 5.88 12 133 9.02 8 8 132 6.06 9 1A1 6.38 9 2 116 1.72 A 1AA 2.78 10 5a 100 5.00 A 133 3.01 11 7 82 8.5A A 100 A.00 12 l 58 1.72 2 86 2.33 13+ 19 860 2.21 23 1,16A 1.98 aFaculty members terminating at the end-of—the tenth year have vested rights. 70 *- r. .‘i! lb ill-.n. 1- ooa>hom q- onswam u m ”ompmo>ncoz u 0 ”oopmo> ”0300000 00 one: mCOHBmSUo COHmmoswop 629* .OH iHH sassy uoIqeaedes 71 As one would expect, the quit-rates for both samples decrease with length of service. If the increasing value of non-vested credits had an inhibiting effect on mobility, the non-vested sample trend should have decreased faster than that of the vested sample and, indeed, it did. This in itself, however, does not prove that the increasing value of non-vested credits was responsible for the dif- ference in the decreasing rates of voluntary separations. Note the wide dispersion of the observed quit-rates for each sample from their respective trend rates. Obviously, factors other than length of service and the increasing value of non-vested credits influenced voluntary separa- tions. It might be possible that the difference in pen- sion plans had no bearing on the difference in trends, i.e., the observations are so widely dispersed that sta- tistically, both samples might have been drawn from the same universe. Using a reliability test for regression coefficients, the probability that both samples are drawn from the same universe is approximately two in ten for both sample regression coefficients.15 Statistically, 15The standard deviation of regression coefficients from one sample to another can be estimated from the following equation: Sb = S y - s y~x S x /N where Sb = the standard deviation of the regression °x coefficient 72 one cannot state that this result validates the hypothesis that credits inhibit mobility; however, it does cast doubt upon a priori judgments that flatly discredit the hypo— thesis. One final observation concerning the data in Table 13 is that voluntary separation rates for the non-vested sample fall off much more sharply from the sixth through the ninth year as compared to the vested sample quit- rates. Since the value of non—vested credits increases with covered service, and the probability of obtaining Sy = the standard deviation of observed values 'X around the regression line S = the standard deviation of the observed X values around the mean N = the number of observations Sby = .35 for the non-vested sample, -x Shy = .32 for the vested sample -x The confidence interval at 1.5 standard errors (including 82 cases out of 100) for regression coefficients obtained from additional samples containing nine observations is -.35 to -1.AO for the non-vested sample and -.86 to +.08 for the vested sample. The regression coefficient for the non-vested sample was -.88, and for the vested sample, -.39. Thus, for each sample, approximately eight of ten sample regression coefficients will fall within a confi- dence interval excluding the regression coefficient of the other sample. Fo; a discussion of the reliability of regression coefficients see: Mordecai Ezekiel and Karl A. Fox, Methods of Correlation and Regression Analysis (3rd ed., York: John Wiley and Sons, Inc., 1963), p. 281. 73 vested status increases with service, the more rapid re- duction in the non-vested quit-rate seems to strengthen the pension-mobility hypothesis. In the previous section, it was presumed that non— vested credit value would not be a significant deterrent to the mobility of employees that started participating in the Illinois Universities Retirement Plan prior to reaching their forties. It was possible to obtain the starting ages of 25 non-vested participants who separated after completing at least five years of service. The ages for the 25 separating participants at the time each entered the plan are depicted in Table 1A. TABLE lA.-—Age at entrance to plan of departing non-vested participants. ’ fi' 7 1* I: ‘cqrfil 1 "fir Age Age Age Age Age 25 29 30 32 36 v 27 29 3O 33 36 28 29 31 33 37 28 29 32 3A 38 28 30 32 35 A5 The mean average entrance age for the 25 partici- pants is 32. Note that only one of the sample of 25 was beyond age A0 at the time of entrance into the plan. The starting ages of those who separated from the vested sam- ple were not available; therefore, a comparison of the two samples cannot be made. In view of the universally 7A acknowledged relationship between mobility and age, one can assume that an inverse quit-age relationship would occur. It seems doubtful, however, that the quit—age relationship alone would be responsible for only one of twenty-five separating participants having begun partici— pation beyond the age of forty. 3': Summary of Results ; _ The step-by-step procedure utilized in testing the pension-mobility hypothesis elucidates the difference i A between what may be termed the "traditional" and the B "appropriate" test. In the "traditional" test, which basically conforms to the type of investigation criti— cally reviewed in the preceding chapter, relevant impedi- ment variables within a non-vested plan are either over- looked or assumed to be of minor consequence. Essen- tially, investigators have homogenized all participants in non-vested plans and have failed to discriminate among participants on the basis of non—vested credits. The "traditional" test is exemplified in the aggre- gate quit—rate comparison as exhibited in Table 11 where no allowance was made for the varying impact that non- vested credits might have on mobility. All participants in the non-vested plan, including those having vested status, were included in the non-vested sample and, there- fore, all were given equal weight in determining the 75 non-vested quit rates. Also, by not isolating the rele- vant variables that determine non-vested credit value, the impact on mobility of factors other than the pension- vesting factor might submerge a genuine pension-mobility relationship. This is exemplified in the quit—rate com- parison illustrated in Table 12 as well as in Table 11. In both tables, factors such as the difference in school prestige and faculty compensation between the samples drown out the capability of determining whether the loss of non-vested credit value might be an authentic impedi- ment to mobility.l6 Thus, in what has been termed the "traditional" approach, investigators have failed to correctly discriminate between non-vested plans and non- /‘ vested credits. In essence, it is not the non-vesting:“/’ 16Of the six universities included in the non-vested sample, the University of Illinois is most comparable in prestige and compensation with the vested universities. According to the AAUP, faculty compensation at the Uni— versity of Illinois was $10,952 in l96A-65, $11,A8O in 1965-66, and $12,132 in 1966-67. Although these values are not as high as the average compensation (including pension plan remuneration values) at the vested schools as exhibited in Table 10, they are far above the average com- pensation for the non—vested schools. The quit-rates dur- ing the non-vested period for the University of Illinois sample were: 6.59 per cent in l96A-65, 5.86 per cent in 1965-66, and 10.00 per cent in 1966-67. The average for the three-year period was 7.55 per cent. Note that these quit-rates are lower than those of the total non-vested sample shown in Table 12 with the exception of the 1966- 67 year. Although the difference in quit-rates between the non-vested sample and that of the University of Illi- nois sample is small, the lower quit-rate average for the University of Illinois sample suggests an inverse rela— tionship between quit-rates and the compensation-prestige variables. / ‘ 76 label on a pension plan that may restrain mobility, but rather the income value lost with separation by partici- pants having non-vested status. In what has been termed the "appropriate" test, recognition is given to the varying impact on mobility that non-vested credits might have on participants within the non-vested plan. Through the utilization of the re- tirement benefit formula coupled with employees' contri- butions, the factors that determined non-vested credit values were isolated. Thus, non-vested service, interest rates, and participants' ages at the time of certifica- tion were found to have an important effect on the value of non-vested credits. The appropriate test, therefore, had to incorporate these factors and, by so doing, the non-vested sample takes on heterogeneous characteristics. Table 13 displays the results of recognizing the impact of length of service on the value of non—vested credits and the concomitant effect on mobility. Al- though the number of quits by year become relatively small in both samples with the approach of vesting, the faster reduction in quit—rates for the non—vested sample gives support, albeit weak, for the hypothesis. When the effect of starting age is recognized (Table 1A), increasing sup- port for the hypothesis is obtained since it was shown that only one of twenty-five non—vested participants having five or more years of service had begun service _._-_. I 77 after the age of thirty-eight. This infers the strong probability that only a very small fraction of those par— ticipants separating from the non-vested plan suffered significant pension credit losses. From this evidence, coupled with the more rapid reduction in quit—rates in the non-vested sample relative to the vested, some sup- port exists for the pension-mobility hypothesis. CHAPTER V SUGGESTIONS FOR FURTHER STUDY Up to this point, the theory of labor mobility with reference to non-vested provisions has been explicated; the literature of this area of economics has been reviewed; and a test of the impact of non-vested credits upon the mobil- ity of college faculty has been conducted. Two significant findings evolved from the test covered in the preceding chapter. First, participant non— *- vested credit income values, and not the act of being vested, were found to be the source of non-vested partici- pant mobility restraint. Second, the vesting and retirement benefit provi— sions of the Illinois plan gave rise to increases in the value of a participant's non—vested credit income with (1) increasing length of service, (2) higher salaries, and (3) higher certification ages. Of particular significance was the strong direct relationship between participant- certification age and non-vested credit income. Given an inverse relationship between age and mobility, it seems probable that relatively few participants in the Illinois 78 79 plan consider separation at a time at which they face serious non—vested credit income losses. Thus, one could not expect a much more rapid decrease in the quit-rates of non-vested faculty, compared to those of vested faculty over the non-vested period, than that exhibited in Figure 1. In fact, a much more rapid increase in the quit—rate differentials would imply that non-vested participants had over—valued their non-vested credit incomes. If all incompletely vested plans were identical in structure to that of the Illinois plan, and if the char- acteristics of all non-vested participants were identical to those in the Illinois sample, then the mobility inhibit— ing impact on the labor force of non-vested credit income would be relatively slight. However, the diversity of in- completely vested plans, and the possible differences in response to non-vested credit income by non-vested em- ployees in such plans from those of employees included in the Illinois sample, preclude direct projection of the re- sults of the Illinois test. One test of a sample of participants in one incom- pletely vested plan, in one occupation and in one industry is not a firm foundation for generalizations concerning the impact of incompletely vested plans on the mobility of all non-vested participants. Diversity of characteristics among plans, coupled with the diversity of the personal characteristics of participants, precludes any such 80 generalization. However, although the test results cannot be projected to all incompletely vested plans, the pro- cedure used to test the mobility impediment strength is sound, i.e., future tests should be directed toward rela- tion of participant non-vested credit income to mobility rather than simply assuming that all participants in a plan are equally inhibited from voluntary separation. FE It would be presumptuous to predict the different possible impacts of identical non-vested credit values on the mobility decisions of workers who are employed in dif— i 1 ferent industries, engaged in different occupations, and possessed of different educations. Only future test re- sults can suggest the relevance of such variables on mobility. However, it is possible to suggest the effect certain provisions of incompletely vested plans may have on the values of non—vested credits and, thus, to evaluate the plan's potential effect on mobility. Plan Differences and Their Impact Upon Mobility of Non-vested Participants Financing, vesting, and retirement benefit formula provisions differ among plans, and must be included in any list of characteristics possessing potential to in- fluence the impediment strength of non-vested credits. Financing Financing provisions are of two types: contributory and non-contributory. The contributory type is exemplified 81 by the Illinois plan, i.e., both the employer and employee share in the cost of the plan. Under a non-contributory plan, the employer bears all of the plan's cost. It can be assumed that participants in contributory plans would be more aware of the existence of the plan than partici- pants in non-contributory plans because of the pension plan cost deductions from their wages. From this it would be expected that contributory plan participants would be more cognizant of pension losses upon separation and, therefore, more inhibited from voluntarily separating from their jobs. However, awareness of losses does not mean that the participant in the contributory plan is more aware of the value of non-vested credit losses, i.e., loss of the employer's contributive share. Furthermore, and probably extremely important in giving rise to a differ- ence in the mobility decision of participants in both types of plans, the non-vested participant in the contri- butory plan receives his contributions (usually with in- terest) upon separation. This might give strong encourage- ment to participants in contributory plans to separate from their jobs,especially to younger participants such as those who left their positions in the Illinois test: Thus, one factor that should be tested is the effect of the type of financing on non-vested participant separation rates. 82 Deferred Vesting Deferred vesting provisions are of two general types, deferred full vesting and deferred graded vesting. Under deferred full vesting, the participant moves from non— vested status to fully vested status upon fulfillment of service and/or age requirements. Under deferred graded vesting, the participant moves from non-vested to fully ”E vested status by stages, e.g., 50 per cent vested after ten years of participation, with increases of 10 per cent per year over the succeeding five years. In both deferred i T full and deferred graded vesting plans, vesting periods vary anywhere from the completion of one year of service to thirty years of service and, in many plans, requires attainment of a minimum age. The predominant service re- quirements among plans are ten and fifteen years, whereas, ages forty and forty-five dominate where minimum age is a requirement.1 Tests could be directed toward ascertaining whether deferred full vesting plans or deferred graded vesting plans that fully vest at the end of the same time period have the greatest effect on mobility restraint. The re- duction in the value of a participant's non—vested credits in deferred graded plans that occurs during the stages of increased vesting should yield greater mobility in such plans as compared to deferred full vesting plans. 1 p. 15. U. S. Department of Labor, Labor Mobility . . ., 83 Of particular importance is the effect of completely non-vested plans on mobility as compared to the effect of deferred vesting plans on mobility. If non-vested parti- cipants are aware of the value of their non—vested credits, then younger non—vested participants belonging to non- vested plans should be more mobile than their counter- parts in plans that defer vesting. For example, assume that a 35-year—old participant has 10 years of service in a deferred full vesting plan that vests after 15 years of service. His counterpart covered by a non-vested plan pay- ing the same benefits at normal retirement age of 65 must maintain his present position for 30 years to obtain re- tirement benefits. The value of non-vested credit income would be higher for the participant in the deferred vest- ing plan because of the higher probability of staying five more years in the plan as compared with 30 more years. In fact, given that mobility decreases with age, it seems probable that firms that have plans that defer vesting have lower labor turnover than plans without vesting. Thus, tests comparing the effect of various types of vest- ing provisions on mobility, might prove that "liberalizing" completely non—vested plans by insertion of deferred vest- ing provisions might induce less overall mobility. Retirement Formulas Retirement benefit formulas fall into three groups: (1) benefits that vary with service and earnings, 8A (2) benefits that vary with length of service, and (3) benefits that are set at a flat rate for all retirees.2 Plans that include earnings in their benefit formulas are of two types, (1) "career average" earnings, the average earnings of the participant over his entire ser- vice, and (2) "final average" earnings, as was the case with the Illinois plan. ? Where benefits vary with service and earnings, the value of a participant's non-vested credit income would be more closely related to the participant's earnings ' than under "flat rate" or "service only" benefit formula plans. Since it is the income differential between the present position and another position that is compared to the value of non-vested credits by the participant, it would seem that non-vested participants with relatively low earnings in plans where benefits are unrelated to earnings would be less mobile than participants in the same plan with relatively high earnings. For example, assume two 50-year-old participants are in the same plan that determines benefits on the basis of length of ser- vice. Further, assume that both have completed 10 years of service, the plan vests upon the completion of 15 years of service, and one employee is earning $5,000 per year 2A fourth type of plan defines contributions but does not define benefits. This type is exemplified by the Teachers Insurance Annuity Association-—College Re- tirement Equity Fund plans that are widespread through— out higher education. The TIAA-CREF plan is discussed below, on p. 89. 85 and the other $10,000 per year. It seems highly probable that the $10,000 per year participant would receive out- side offers whose income differential would be, in abso- lute dollars, much higher than that received by the em- ployee earning $5,000 per year. However, the value of non-vested credits to each participant would be identi- cal. Thus, the higher paid participant has less effec- tive mobility restraint than the lower paid participant. Given that higher paid employees have greater value to the firm, probably because of higher skills, such plans would impede the mobility of the less skilled employees more than those with greater skills. If one of the pur- poses of incomplely vested plans is to "lock-in" skilled employees, the plans varying benefits with earnings should be more successful than those plans omitting earn— ings as a determinant of retirement benefits. Quit-rate comparison tests of samples of such plans would be neces- sary to confirm or reject this hypothesis. In summary, financing, vesting, and retirement benefit provisions in incompletely vested plans influ- ence the value of a participant's non-vested credits and, thus, determine the magnitude of the concomitant restraint on labor mobility. Because of the variation in such pro- visions among plans, it is impossible to project the re— sults of a test of a single plan's impact on labor mobil- ity to all incompletely vested plans. 86 Mobility Implications of Vesting When an employee is separated with a vested right, the amount of the benefit he will receive at retirement is fixed at the time of separation.3 Thus, the separated vested employee receives no return on the value of his vested credits to retirement, nor does he receive the benefit of any improvements in the benefit formula that might occur between separation and retirement. Further- more, given rising wage levels, especially evident with inflation, plans that relate benefits to earnings and length of service, gain mobility impediment strength with- out a change in the benefit formula. Consider the following example for a hypothetical employee in the Illinois plan. Assume the employee begins participation in the Illinois plan at age 30, works for 15 years, quits, accepts another job at which he works for 15 years and is covered by a plan having a retirement formula identical to that of the Illinois plan. Further, assume that his "final average earnings" in his first position is $6,000 per year, and in the second, $12,000 per year. Given the same retirement formula in both plans, his re- tirement income at age 60 is: (1 2/3 percent x $6,000 x 15/12) + (1 2/3 per cent x $12,000 x 15/12), or $375 per month. Assuming his "final rate of earnings" would equal $12,000 per year if he had stayed at Illinois, his monthly 3Bernstein, The Future of Private Pensions, p. 258. 87 retirement income would have been: (1 2/3 per cent x $12,000 x 30/12), or $500 per month. Thus, in plans that determine benefits on the basis of length of service and earnings, the value of a participant's vested credits in- creases with increases in earnings. A further problem involves the possible effect upon mobility of workers with vested rights in plans that specify retirement benefits based upon length of service when such plans have experienced periodic improvements in their benefit formulas.“ For example, the following is the history of benefits under the Caterpillar Tractor Cor- poration--United Auto Workers non-contributory plan that was initiated on December 1, 1950.5 December 1, 1950; $1.50 per month times years of service August 8, 1955; $2.25 per month times years of service October 1, 1958; $2.50 - $2.80* per month times years of service October 1, 196A; $3.95 - $A.25* per month times years of service October 1, 1965; $6.00 per month times years of service October 1, 1967; $7.00 per month times years of service October 1, 1968; $7.00 - $8.00* per month times years of service. *The higher figure is applicable to specific skill categories, e.g., toolmakers. “Of 100 negotiated plans covering 3,508 thousand workers surveyed by the Department of Labor, A7 plans covering 2,065 thousand participants improved benefits during the period, 1961-196A. Of the A7 plans experienc- ing improvement, 17 plans covering 895 thousand workers had improvements in excess of 50 per cent. See, Harry E. Davis, "Changes in Negotiated Pension Plans Under Collec- tive Bargaining, l96l—6A," Monthly Labor Review (October, 1965), p. 1217. 5Benefit data obtained in an interview with Smith Applegate, Insurance Department, Caterpillar Tractor Corp., Peoria, Illinois, December 9, 1968. 88 The plan fully vests benefits upon the completion of 10 years of service and begins monthly annuity payments to participants retiring at age 62. If a participant in a skilled classification had separated in September, 196A, he would receive $28 per month throughout retirement. If he had stayed on the job to October 1, 1968, he would re- ceive $8.00 per month for each of his 1A years of service, or $112 per month throughout retirement.6 Furthermore, those participants who retire directly from employment at Caterpillar receive all succeeding general benefit im- provements that occur during their retirement years, whereas the separated vested employee's monthly retirement benefits remain fixed throughout retirement. Again, as was the case with service-earnings benefit formula plans, the terminated vested employee might obtain employment covered by another plan and receive retirement benefits from that plan as well as from the first plan. However, it is doubtful whether the combined benefits would surpass the benefits that he would have received had he stayed in his first job. Since benefits are not related to service in plans that specify flat-rate benefits to all employees with a vested right, this type of plan does not offer the same 6Where plans are negotiated and unions are pressing for pension benefit improvements, participants might be- come highly immobile as the contract deadline is approached. 89 type of restraint upon vested employees as do the service benefit formula plans. Finally, most plans specify a minimum service re- quirement, usually 10 or more years, to obtain retirement benefits.7 Thus, older employees, having obtained a vested right, and faced with the possibility of not being able to make up a portion of their "lost vested credits" in another plan, might prove to be the most "locked-in" of all vested participants. In summary, investigations of the mobility impact upon participants with vested rights, in plans yielding increasing vested credit values with improvements in ser- vice formulas, might find such plans to be a major detri- ment to labor mobility. Conclusion The only type of pension plan that seems not to contain inherent labor mobility restrictions is that which treats pension credits as current remuneration. Examples of such plans are those contracted with the Teachers In- surance Annuity Association-~College Retirement Equities Fund. These plans formalize the contributions to be made by the employer (and by the employee if the plan is con— tributory), and omit a benefit schedule. Such plans are termed "defined contribution plans." All contributions 7U. S. Department of Labor, Private Pension Plan Benefits, p. 7. :~ 31;, \ 90 are fully and immediately vested, are expressed as a per- centage of the employee's earnings, are deposited with TIAA-CREF in exchange for "accumulation units," and each employee has a separate account with TIAA-CREF. Up to 75 per cent of the contributions may be directed by the em- ployee to be used to purchase CREF accumulation units, and the remainder is used to purchase TIAA accumulation units. TIAA places its funds in debt instruments, while CREF's funds are placed in equity instruments. All in- terest and dividents from the funds' portfolios are appor- tioned to each participant's account, according to the number of accumulation units owned by the participant. The dividends, interest, and contributions are used for the purpose of increasing the amount of accumulation units. Therefore, the value of each unit depends upon the market value of the TIAA-CHEF portfolios. Upon retirement, the value of the participant's accumulation units are the basis for his retirement annuity.8 Note that vesting, financing, and retirement bene- fit provisions in TIAA-CREF plans impart no separation restraint on the participant. By relating contributions to current earnings and by fully and immediately vesting such contributions, the separating participant suffers no loss of retirement benefits, i.e., the value of his 8College Retirement Equities Fund, Rules for Deter- mining Benefits (New York: CREF, 1960). 91 pension credits earned up to the time he separates, will have the same value at retirement as if he had not sepa- rated. In terms of the possible effect on labor mobility, and the concomitant effect on resource allocation, only those plans which fully and immediately vest current con- tributions, free the participant from suffering possible pension losses due to termination. By isolating the sources of pension plan mobility restraint (non-vested credit income and potential vested credit income losses), investigators can more accurately measure the mobility impediment effect of pension plans for labor. SELECTED BIBLIOGRAPHY SELECTED BIBLIOGRAPHY Books Balassa, Bela. "Success Criteria for Economic Systems." Comparative Economic Systems. Edited by Morris Bornstein. Homewood, Illinois: Richard D. Irvin, Inc., 1965. Bernstein, Merton C. The Future of Private Pensions. London: The Free Press 6? Glencoe, l96A. Brown, David G. The Market for College Teachggg. Chapel Hill, North Carolina: The University of North Caro- lina Press, 1965. Boulding, Kenneth E. Economic Aanlysis. 3rd ed. New York: Harper and Brothers, 1955. Carter, Allan M., and Marshall, F. Ray. Labor Economics. Homewood, Illinois: Richard D. Irwin, Inc.,71967? Chamberlain, Neil W. The Labor Sector. New York: McGraw- Hill Book Company, 1965. Ezekial, Mordecai, and Fox, Karl A. Methods of Correlation and Regression Analysis. 3rd edT'iNew York: John Wiley and Sdns,vlnc.,'l963. Greenough, William C., and King, Francis. Retirement and Insurance Plans in American Colleges. New York: Columbia University Press, 1959. Holland, Daniel. Private Pension Funds: Prgjected Growth. Neg York: National Bureau of Economic Research, 19 6. Kerr, Clark. "Social and Economic Consequences of the Pension Drive." National Industrial Conference Board, Studies in Personnel Policy_No. 103. New York: National Industrial Conference Board, 1950. 93 9A Lester, Richard A. Lester. Economics of Labor. 2d ed. New York: Macmillan Company, l96A. Marshall, Howard D. The Mobility of College Faculties. New York: Pageant Press, Inc., l96A. McGill, Dan M. Fulfilling Pension Expectations. Home- wood, Illinois: Richard D. Irwin, Inc., 1962. McGill, Dan M. Fundamentals of Private Pensions. Home- wood, Illinois: Richard D. Irwin, Inc., 196A. Palmer, Gladys. Labor Mobility in Six Cities. New York: Social Science Research Council, 195A. Rubin, Harold. Pensions and Employee Mobility in the Public Service. New York: The Twentieth Century Fund, 1965. Wight, Bakke E., et a1. Labor Mobility and Economic Opportunity. New York: John Wiley and Sons, Inc., 195A. Government Davis, Harry E. "Changes in Negotiated Pension Plans Under Collective Bargaining, l961-6A." Monthly Labor Review (October, 1965), pp. A5-A9. U. S. Congress. Joint Economic Committee. Private Pen- sion Plans. Hearin s before a subcommittee on Fiscal Policy, Part , 89th Cong., 2d sess., 1966. U. S. Department of Labor. Bureau of Labor Statistics. Digest of 50 Selected Pension Plans for Salaried Employees, Summer, 1965. Bulletin Number 1A77. Waggington, D.C.: Government Printing Office, 19 . U. S. Department of Labor. Bureau of Labor Statistics. Digest of 100 Selected Pension Plans Under Collec- tive Bargaining, Late l96A. Bulletin Number 1A35. Wazhington, D.C.: Government Printing Office, 19 5. U. S. Department of Labor. Bureau of Labor Statistics. Labor Mobility and Private Pension Plans. Bulletin Number 1A07. Washington, D.C.: Government Print— ing Office, l96A. 95 U. S. Department of Labor. Bureau of Labor Statistics. Private Pension Plan Benefits. Bulletin Number 1A85. Washington, D.C.: Government Printing Office, 1966. «101111121 Brissenden, P. F. "Labor Mobility and Employee Benefits." Labor Law Journal (November, 1955), 762-768. Greene, Mark. "Fringe Benefits or Salary." Journal of Marketing (October, 1963), 63-68. 0 - = ‘v‘ "x I.‘ ._ . ‘3' Lfimik ' Lester, Richard A. "Benefits as a Preferred Form of Compensation." Southern Economic Journal (April, 1967), A88-A95. Lurie, Melvin. "The Effect of Non-vested Pensions on Mobility: a Study of the Higher Education Indus- try." Industrial and Labor Relations Review (January, 1965), 22A-237. Parnes, Herbert. "Workers Attitudes to Job Changing." The Reluctant Job Changer. Edited by Gladys L. Palmer. Philadelphia: University of Pennsylvania Press, 1962. Ross, Arthur M. "Do We Have a New Industrial Feudalism." American Economic Review (December, 1958), 903-920. Tilove, Robert. "Social and Economic Implications of Private Pensions." Industrial and Labor Relations Review (October, 1960), 2A-3A. Wagner, Ludwig A., and Bakerman, Theodore. "Wage Earners' Opinions of Insurance Fringe Benefits." The Journal of Insurance, 27 (June, 1960), 17-28. Pamphlets College Retirement Equities Fund. Rules for Determining Benefits. (New York: CREF, 1960). State Universities Retirement System of Illinois. Hand- book of Information, State Universities Retirement System. Urbana: 1965. M71111711777leMlflffllflflljlflifflflil“