SENATE Report of the Ad Hoc Committee on Benefits Redesign

Benefits Redesign at the University of Pennsylvania
A Review and Analysis of the Recommendations Published on
February 11, 1997, by the Benefits Advisory Committee

Submitted to SEC March 13, 1997


The Faculty Senate ad hoc Committee to Review the Benefits Redesign Report was appointed by the Chair of Faculty Senate, Peter J. Kuriloff. The Committee was charged with the limited task of reviewing and analyzing recommendations made by the Benefits Advisory Committee (Almanac, February 11, 1997) for changes in four components of the current benefits program that impact directly on the welfare of members of Penn's Faculty Senate. In view of this specific charge, the ad hoc Committee did not consider other components of the entire benefits program or the relationship between the benefits and salary components of total compensation. Accordingly, the review and analysis reported here only addresses changes recommended by the Benefits Advisory Committee in (a) health care insurance, (b) graduate tuition remission for spouses and dependent children of Penn employees (c) life insurance, and (d) benefits for regular part-time employees.

Since receiving an advance copy of recommendations for benefits redesign on February 5, 1997, the ad hoc Committee met several times during the past month to review and analyze these recommendations. During the course of this review, the Committee requested, and received, additional information from the Office of the Provost and the Division of Human Resources in its efforts to understand information and principles used by the Benefits Advisory Committee in forming its recommendations. Nonetheless, the work of the ad hoc Committee was constrained by the short period of time and incomplete detailed information available to perform its task. The content of this report, therefore, is necessarily limited by these conditions under which the Committee was obligated to function.

The ad hoc Committee performed its review and analysis of recommended changes in selected benefits from the perspective of the faculty as a whole. Of course, the Committee understood that the changes as proposed will have differential effects on individual faculty members, but it judged that the general welfare of the members of Faculty Senate should be its overriding concern.

Health Care Insurance

The primary rationale for the University's benefits redesign is a contention that, compared to other employers with whom Penn competes in various labor markets, Penn's compensation package is overly rich in benefits.

The health benefits redesign therefore involves two changes: 1) a modest alteration in the set of health plans among which employees can choose and 2) a substantial increase in the fraction of the cost of these plans to be paid by employees. This increase in employee contribution is not distributed evenly across plans.

Compared to the current set of health plan offerings, two plans which were chosen by a modest number of employees were deleted: the higher deductible "comprehensive" indemnity plan and some of the HMO options, most importantly the HMO offered by QualMed. A new option, an HMO with a Point of Service option supplied by Keystone Health Plan East, has been introduced.

The fraction of the cost (premiums for outside plans and benefits paid for self-insured plans) expected to be paid by employees has been increased from the 11% level which prevailed in FY 1997 to the level of 17%, approximately equal to the level in FY 1994. Some plan options experienced very modest increases in premiums. The lower deductible indemnity "Plan 100" will have increases in employee contributions of only a few dollars a month. The Keystone and US Healthcare HMOs, which had zero employee premiums in 1997, will require annual premiums of $120 for individuals and $312 for families. The redesign plan document claims that these increases are virtually identical to the expected cost for adding a prescription drug benefit that was formerly lacking only from this option.

The largest increases in employee paid premiums will come from people who chose any plan but Plan 100 in 1997 and who in 1998 chose either the PENNCare plan or the new Keystone Point of Service (POS) plan. All plans except Plan 100 had virtually zero premiums in 1997, while the annual employee premiums for 1998 are proposed to be set at $240 for individual coverage and $624 for family coverage for the Keystone POS option, and at exactly twice these levels for the PENNCare option.

Because the consultants calculated the premium costs in each year using two entirely different methods, it is impossible to determine precisely the extent to which any higher employee premiums match higher expected benefits for each of the plans. The annual increase in total health care benefits costs is expected to be very modest (almost 5%), while the employee premium share is forecasted to increase by about 50%, so one must conclude that the increase in employee premiums must be considerably greater than the increase in expected benefits or costs. Beyond this conclusion, the following discussion of the plan-level changes must be regarded as somewhat speculative. Nevertheless, it does appear, based both on data in the description of the plan changes and statements by the Administration, that the proposed new premiums for Plan 100 and the USHC/Keystone HMO reflect almost none of the increased premiums for current benefits. That is, after adjusting for the cost of the prescription drug benefits added to the HMO plan, employees who formerly chose these plans and stay in them are bearing little or none of the increased employee premiums. Instead, those who pay additional premiums of from $240 (individual, POS) to $1,248 (family, PENNCare) bear the great bulk of the burden of the additional employee premiums.

There appear to be two major problematic features of this benefits redesign. Employees are being asked to bear more of the cost of their benefits without receiving any increase in money wages as an offset, and the burden of that cost is distributed very unevenly across employees depending on their benefits choices. It appears that those employees who formerly chose the comprehensive or PENNCare options bear all of the cost of the increased employee contributions in excess of any increased anticipated total costs to the University. A third issue is whether the proposed blending of plan costs is the appropriate pricing strategy or whether each plan option should be self-supporting.

Increases in employee premium contributions have the same effect on gross pay as reductions in money wages. Because employee premium contributions are tax shielded, the two are not exactly equivalent, but it is clear that when employee payments for health benefits are increased in excess of any increased cost of those benefits, the value of total compensation has been reduced. Other things equal, employees may be expected to oppose reductions in the value of their total compensation. Had the increased employee contributions been distributed approximately evenly across employees, it would be possible to offset these higher contributions by an increase in the money wage increases in 1998 in excess of the raises that would otherwise have been paid. While it is still possible to provide such offsets, they will overcompensate those who remain in Plan 100 and the US Healthcare/Keystone HMO, and under compensate those who were in the Comprehensive or PENNCare plans and who do not switch to the HMO or Plan 100.


The primary problem in Penn's current health benefits offering is that, except for Plan 100, all other offerings carry approximately zero employee premiums. This is alleged to result in distorted choices across plans (since employees have no incentive to take into account the difference in costs to the University for different plans) and a compensation package which induces employees with working spouses to elect family coverage at Penn, rather than at the spouse's employer. This situation arose because Penn's employer contribution policy was based on a fraction of the cost for Plan 100; as premiums for that plan have risen relative to other premiums in recent years, the employer's contribution has exceeded the total premium for the other insurance options, resulting in zero or in some case negative premiums for them. The redesign is intended to correct this design flaw. However, one may argue that these negative contributions can serve as an incentive encouraging employees into the most cost-effective plans.

While acknowledging that Penn has experienced the same problem in its benefits design as have other employers (including the Federal government and the large California Public Employees system, two common benchmarks), the proposed solution is open to question. The reasons for skepticism are indicated in the description above: the increased employee contributions to correct the acknowledged design flaw are not distributed evenly across employees or in proportion to a health plan's cost to the University, and are not proposed to be offset by higher money wage increases.

With regard to the issue of money wages, the argument made in the description of the proposed plan for raising the average employee contribution and not promising a money wage offset is only that the proposed employee contribution level of 17% is "still well below the average in the private sector," with no mention of comparative wage levels. If the problem with Penn's current package is that it is too benefits-rich and cash-poor, the solution would seem to involve both a reduction in University benefits contributions and an increase (compared to what otherwise would have occurred) in money wages. It seems likely that the high increases in University contributions which occurred in FY 1995-1997 were offset by lower increases in money wages (compared to what would have occurred if benefits costs had increased less rapidly), so that attempting to correct matters by actually raising the employee payment for benefits, recapturing some of the excess contribution of previous years, without returning some of the wage offset, will not in fact redress the balance. The alternative hypothesis is that during the period of excessive benefits contribution money wages were for some reason not reduced in an offsetting fashion; this would lead to the conclusion that Penn employees are currently overcompensated in total relative to the labor market. No evidence was provided that would support this contention; indeed, the assertion in the introduction that "the cost of employee benefits is negatively affecting Penn's ability to ... (achieve) competitive salary for faculty and staff" implies that it is not so.

The more controversial issue concerns the uneven pattern of employee premium increases. While nowhere stated in the explanatory documents, the University appears to have chosen to keep the employee premium contributions for the HMOs low because these plans are disproportionately (though not exclusively) chosen by low wage employees. It is less obvious why it has chosen to keep the Plan 100 premiums low, since those plans are disproportionately chosen by higher salaried faculty. It has been suggested that, if the University wishes to cushion the blow of higher employee premium contributions on lower wage employees, a preferable way to do so would be to raise the money wages of those employees--since this would avoid "subsidizing" the employees who chose the HMO option but are not low-wage. Others have questioned the desirability or the feasibility of using compensation policy to redistribute income among employees when compensation policy is also supposed to be competitive at all wage levels.

Owing both to the complexity of the problem and the little time available to secure needed information for this review, no specific alternative for redesign of health benefits can be suggested at this time. The objective of correcting the design flaw associated with the former contribution policy is a worthy one, but those design flaws need to be clarified and verified to a greater extent to assure that the new design does not itself contain implicit flaws. In addition to the concerns already mentioned, there was fear that the proposed modest increases in employee contributions to Plan 100 premiums cannot be sustained, if the costs of that plan continue to rise at both its historical and projected rates. Paying more attention to wage offsets overall and to real economics for health plans ought to guide the "redesign of the redesign" which we favor. Additionally, there should be reconsideration of the contents of the various plans as well as their pricing, especially in light of concerns that have been raised about the adequacy of Penn's plans in respect to mental health benefits, nursing home care, home health care and the size of the lifetime cap on benefits. Long-term care and quality of care issues also should be considered in a potential redesign. In the meantime, and probably for FY 1998, the proposed moratorium by SEC on changes should probably apply to health benefits, unless and until a more satisfactory alternative can be developed. In the absence of a complete moratorium on change, perhaps a less dramatic movement in plan pricing could occur in FY 1998.

Graduate Tuition for Spouses and Children

The rationale for the elimination of the graduate tuition benefit for spouses and dependent children is that this is a costly benefit ($1.3 million a year) that is used by a very small percentage of Penn employees at any one time and exceeds what is offered by peer institutions. On the basis of the additional information that we received, we conclude, however, that the actual net cost of this benefit to the University is considerably less than what is claimed and we note that, while the number of employees using it at any one time is small, the percentage who use it at some point over the course of their entire career at Penn may be considerable. Furthermore, eliminating this benefit has a high price in the lost trust of those employees who came to Penn in the expectation that it would be available to them. In addition, there is an advantage for recruitment and retention of the faculty in a relatively inexpensive benefit that outstrips what peer institutions offer.

We suggest that the existing benefit be grandfathered for all employees arriving at Penn through January 1998, and not available for employees hired after that date. A further possibility is that the benefit could be made available to employees arriving after January 1998 in a somewhat reduced and streamlined form, namely an across-the-board benefit of 50% of tuition for spouses and dependent children alike (instead of 50% for spouses, and either 75% or 100% for children depending on the school), although the resulting savings would be small.

Life Insurance

Under the current life insurance program, faculty and staff receive flex credits equivalent to purchase a multiple of 1 to 4 times salary depending upon their age. The age brackets are as follows:

     Under 36     4.0 X benefits base
     36 to 45     3.5 X benefits base
     46 to 50     3.0 X benefits base
     51 to 64     2.5 X benefits base
     65 to 70     2.0 X benefits base
     Over 70      1.0 X benefits base

Once Flex Credit allotments are received, staff may choose to purchase $50,000 in coverage, or any multiple of salary from 1 X to 5 X salary provided this multiple does not exceed the cap of $300,000. There is a core amount of required insurance coverage which is the lesser of $50,000 or 1 X benefits base.

The proposed changes to the University's group life insurance program are designed to move to a system whereby the University funds a life insurance benefit of one times salary for employees with the balance of Flex Dollars providing a one-time increase in base salary. This change to the group life insurance program basically eliminates the PennFlex (flexible benefits) program. An individual may also purchase up to $500,000 at existing age-related rates during fiscal year 1997-1998. Future rates most likely will depend on plan experience so that it is very difficult to assess the long-term effect of the proposed change.

The overall implications of this change for individual faculty members are difficult to assess because there is so much individual variation in the benefit. The current program provides a level of coverage determined as a varying multiple of salary based on a participant's age. Age also determines the cost per thousand dollars of life insurance. The multiplication of the age cost of the coverage determines the Flex Dollars that can be used to purchase life insurance or spend on other PennFlex benefits. The proposed recommendation suggests that this benefit could be purchased with after-tax dollars, saving participants a portion of the tax associated with imputed income on group life insurance coverage over $50,000. This saving would occur if the premium rates at which an individual purchases insurance in the University group are less than the rates on which imputed taxable income is based. It is not clear that everyone would be better off with this change, although many participants may well be advantaged by the tax effect. A lower paid employee not subject to taxes may be marginally worse off. Also, the change calls for a one-time salary adjustment giving participants the dollars used to fund their life insurance. Under the present system, the Flex Dollars increase with age. A participant close to retirement and about to cross an insurance price threshold may be better off under the old system. However, there are countervailing beneficial effects. A participant may have a contribution to their retirement plan increased by the proposed one-time change in salary.

As the preceding paragraph indicates, detailed actuarial computations would be necessary to judge whether the community is better or worse off. Introducing the time element in terms of how long one continues working and the pension impact makes the problem particularly cumbersome. In all likelihood, there are some beneficially impacted and some who are marginally worse off.

There are some other policy implications of the change in the life insurance program. It would eliminate the Flex Credit pool associated with the PennFlex program. The Flex Credit was designed to manage benefits for a diverse workforce since multiple benefit programs can be offered and contributions can be made to a credit pool. Employees then pick and choose which benefits they want and where they will spend their Flex Dollars. A decision on whether to disassemble the flex benefit mechanism should be made in connection with a long-range plan. Some consideration should be given to the efficiency of a credit pool, how benefits will be managed in the future, and how new benefit options will be added and paid for. If the proposed change is enacted, it is recommended that a new more inclusive flexible benefits program be studied for possible future implementation.

It is recommended that the amount of any salary increase awarded because of the additional Flex Dollars added to base salary be reported to individual employees as a specific percentage of base salary along with the specific percentage increase in base salary (if any) awarded as an annual increment.

Benefits for Regular Part-Time Employees

In the judgment of the ad hoc Committee, the addition of the option of a Health Care Pre-Tax Account for regular part-time employees improves significantly the very limited benefits offered to this group. Under current circumstances, restricting this account to a maximum of $1000 appears to be reasonable.

However, the benefits provided for regular part-time employees should be given further consideration. Within this group of part-time employees are some who no doubt need and have a justifiable claim to more generous treatment, while there are others who have little need or justification for more generous treatment. For example, the Committee has heard of couples who share a full-time position as their primary source of employment and who are long-term employees in vital and valued positions. By contrast, the Committee has also heard of others who hold minor part-time appointments in marginal positions and who have full-time employment (and associated benefits) elsewhere. Consideration should be given to developing principles that can be used to identify subsets of regular part-time employees for whom more generous benefits should be provided.


In conclusion, the members of the ad hoc Committee do not believe that Penn has a "benefits philosophy," or, even more important, a "total compensation philosophy." By this we mean a set of long-term principles that can be used in the design of periodic changes to the benefits program and salary levels. By contrast, the current set of recommendations in selected benefits, as published in Almanac, certainly is not framed by a stated philosophy, and no such philosophy has been made public. One apparent consequence of the lack of a total compensation philosophy is the use of the piecemeal strategy for making changes in the components of a benefits program (as represented by the current round of recommended changes). Next year, this strategy will again be used in a scheduled review of retirement and disability benefits. The ad hoc Committee fundamentally disagrees with, and disapproves of, the piecemeal strategy because of two main problems with it: (a) changes in one benefit may impact on one or more other benefits not under review, and (b) reductions in benefits without offsetting adjustments to salary may well reduce total compensation.

The ad hoc Committee recommends that a Penn "total compensation philosophy" be developed and published, and that changes to benefits and salary be made in accordance with this philosophy. In the view of the Committee, such a philosophy should embrace the following elements as a minimum:

Faculty Senate Ad Hoc Committee to Review the Benefits Redesign Report

Erling E. Boe, Professor of Education
Sheila H. Murnaghan, Associate Professor of Classical Studies
Mark V. Pauly, Bendheim Professor of Health Care Systems, Public Management, Insurance, and Economics
Jerry S. Rosenbloom, Frederick H. Ecker Professor of Insurance and Risk Management

Response of the Associate Provost and Deputy Provost.


Volume 43 Number 27
March 25, 1997

Return to Almanac's homepage.

Return to index for this issue.