As an interim topic, the Legislative Commission on Pensions and Retirement chose to review service credit purchase provisions added to statutes in 1999, 2000, and 2001 and the procedure in law to compute the purchase prices to receive service credit under these provisions. The pricing procedure in law, which is an attempt to estimate the full actuarial value of the service credit purchase, is used to determine the price to be charged for the service credit being purchased. The procedure currently used was enacted in 1998 with an expiration date. The various statutory service credit purchase provisions which use that pricing methodology were enacted shortly thereafter and were also viewed as temporary, with expiration dates specified in the enacting legislation. As the expiration dates came near, supporters of these provisions urged an extension or legislative action to make the provisions permanent. The Commission and the Legislature, lacking sufficient time for a detailed study of the impact of these provisions on the purchasers and pension funds, extended the dates on a few occasions but did not make the provisions permanent. The provisions were most recently set to expire in 2003, generally in May, but the 2003 Legislature extended the provisions for one more year, to 2004.

The 2004 Legislature faces the issue of what action to take regarding these service credit purchase provisions and the accompanying method now in law to compute full actuarial value. The first consideration memo provided background by reviewing the Commission’s policy document regarding service credit purchases, by reviewing the general law service credit purchase provisions added to statutes in recent years which are set to expire, and by reviewing the method in law (Minnesota Statutes, Section 356.55) used to compute the purchase price of these service credit. The second consideration memo noted numerous conflicts between the Commission’s policy statement and the various service credit purchase provisions, and reviewed the available data on full actuarial value purchases. That memo raised considerable reservations about the accuracy of the current full actuarial value estimation method. One conclusion is certain; the prices that individuals paid to receive service credit is not at all consistent with the liabilities created in the pension plans due to the purchases as reflected in the plan actuarial reports following those purchases. This third consideration memo attempts to lay out the various options available to the Commission and reviews the implications of these options.

The broad issue now facing the Legislature is what to do about the method in current law to compute the prices to be charged in these full actuarial value service credit provisions, and what to do with the service credit purchase provisions themselves, since the current methodology and the provisions are set to expire in May or June of 2004.

The second consideration memo reviewed the service credit purchases reported in the 2000 through 2002 actuarial valuations and compared those prices, which are intended to cover the liabilities being created, with the liabilities reflected in the actuarial studies due to those purchases. The material raised considerable reservations about whether the method in use is sufficiently accurate, and more generally, whether any method can be developed which is sufficiently accurate to permit service credit purchases to be encouraged and widely used.

The questions regarding the current full actuarial value methodology are whether to allow the methodology to expire, to extend it, to extend it with revision, or to make it permanent with or without revision. Regarding the service credit purchase provisions, the Legislature could allow all of these provisions to expire, it could extend some or all of them with revision, or it could make some or all of them permanent with or without revision.

The methodology issue and the service credit purchase provision issue are not fully separate. To an extent, the Commission and the Legislature can revise the full actuarial value methodology by revising it or by replacing it, but the Legislature can never be fully confident that the methodology, however revised, will provide good estimates in all situations. To minimize harm or the potential for harm to individuals purchasing service credit or to the pension fund providing the coverage, the Legislature may choose to minimize its use by allowing some or all of the service credit purchase provisions to expire. The Legislature may also wish to revise any remaining service credit purchase authority by reducing the total amount of service that may be purchased or by other revisions to more narrowly restrict eligibility.

This section presents several potential courses of action and notes the implications of the actions. The Commission may also wish to consider several of these actions in combination. Other alternatives not covered here are also possible.

1. Retain the Current Full Actuarial Value Methodology Without Revision, Either by Another Temporary Extension or by Making the Provision (Minnesota Statutes, Section 356.55) Permanent. After the revised full actuarial value estimation method (Minnesota Statutes, Section 356.55) was enacted in 1998, numerous full actuarial value general law service credit provisions soon followed. Many teachers and other public employees have used the provisions. For fiscal years 2000 through 2002, the years for which data are available, 725 Teacher Retirement Association (TRA) members purchased service credit under these provisions using the new methodology. In all other plans combined, there were another 230 purchases.

   One reason to continue the current procedure is that the new methodology has been popular with teachers and, to a lesser extent, with other public employees. Over 700 teachers have used the provisions to purchase service credit, permitting them to retire earlier under the Rule-of-90 or other early retirement provision, or at least to increase the monthly retirement benefit the teacher will receive.

   There are at least three reasons for not extending the 1998 methodology or not making that methodology permanent. At a minimum, the method needs fine-tuning. Differences between the assumptions used in the full actuarial value computation and the corresponding assumptions used in the actuarial reports for the plan that covers the member need to be reviewed. An earlier staff memo noted that the 1998 methodology assumes five percent salary increases until retirement, regardless of which plan covers the individual. Most plans are now using a select-and-ultimate salary increase assumption in the plan’s actuarial work, and that assumption rarely predicts a five percent yearly increase. The 1998 methodology also assumes an 8.5 percent investment return on the money received from the service credit purchase until retirement. It has been noted previously that an 8.5 percent return may not be obtainable in the marketplace when the service credit is purchased. Further, the Minneapolis Employees Retirement Fund (MERF) uses a 6.0 percent investment return assumption in its actuarial work, not 8.5 percent, and there have been several service credit purchases in MERF authorized by the Legislature in recent years. Lacking a clear justification that the assumptions used in the 1998 methodology are reasonable for the individuals or situations where service credit is purchased, it would be preferable to use the same assumptions as those used in the actuarial reports for the applicable plan.

   The second concern is lack of agreement between the price of service credit under the 1998 methodology and the liabilities that are created in the pension fund due to those purchases, as reflected in the next actuarial valuation. The second consideration memo noted that in the over 900 cases in the study, the price an individual paid for the service credit purchase never matched the liability indicated in the plan’s actuarial report following that purchase. These differences would not be troubling if they tended to be minimal. What is troubling is the size of the differences. The differences tend to be very large, many in the tens of thousands of dollar ranges. If this review is credible, it indicates that considerable harm may be created by these purchases. In some cases, the fund is harmed; in other cases it is the individual purchasing the service credit that may be harmed.

   Arguments can be made that the prices produced by the 1998 methodology are fair; the 1998 methodology and the actuarial valuation results reflect different concepts. However, the data collection requirement included in the 1998 methodology legislation was added for the explicit purpose of determining whether the methodology was computing correct full actuarial value estimates rather than creating windfalls and harm. If that review cannot be relied upon to provide that determination, then the Legislature is still casting about in the dark on this matter. Failure to prove that considerable harm is occurring seems a shaky foundation for a recommendation to continue use of the current method on either a temporary or a permanent basis. A further problem is that in the future new data will not be available because the data collection and processing was eliminated from the Commission-retained actuary’s contract in a budget-cutting move.

   The actuary retained by the Commission developed the current methodology, but while working on behalf of pension fund administrators rather than on behalf of the Commission. Pension fund executive directors work for boards on which employee groups are heavily represented. Some of these employee groups favor service credit purchases and may favor subsidized purchases. To the extent that the objectives of the Commission and the objectives of the pension fund administrators may not fully align on this issue, and given that the Commission-retained actuary was retained by both organizations, the actuary is placed in a difficult position. The actuary’s objectivity may be perceived as being compromised.

   A third justification for not continuing to use the 1998 procedure is the strong support for this method by teachers and some other public employees, which suggests that teachers and other public employees believe these purchases provide the individual with benefits that exceed the cost. If they are correct, the purchases generally harm the pension fund. It would not be worth fighting for a process that provides no net value.

2. Allow the 1998 Full Actuarial Value Methodology to Expire. Under existing law, if the 1998 full actuarial value methodology (Section 356.55) is allowed to expire, the full actuarial value estimation procedure previously used (Section 356.551) will again become effective. The pre-1998 estimation method was criticized in the past by some pension fund administrators as producing estimates which are too high, particularly for individuals who were very close to qualifying for the Rule-of-90 without the purchase. However, the study reviewed in the second consideration memo found that under the 1998 procedure, individuals who purchased service credit and then retired before the next actuarial valuation tended to receive windfalls, often of considerable size. That result suggests that if there was a problem with the pre-1998 method, the 1998 method was a considerable over-correction. Permitting Section 356.551 to become effective seems appropriate if the Commission concludes that the pre-1998 method is the more accurate of the two, or if the Commission concludes that service credit purchases should be discouraged compared to the current situation to ensure that the pension fund is not harmed.

3. Retain the Current Method with Revisions. The Commission may wish to retain the current method, but with revisions to minimize its shortcomings. Possible revisions include the following:

   1. Revise the computation to use the investment return, mortality, and salary increase assumptions applicable to the pension plan in which the individual is purchasing service credit. This would at least establish consistency between the calculation and the assumptions that the actuary contends should apply to the plan as a whole.

   2. As an alternative to #1, use investment return, mortality, and salary increase assumptions specifically intended to protect the pension fund.

   The concern is that the individual and the situation may not be typical, leading to subsidies if the "average" assumptions are used. Using the assumptions applicable to the plan, as suggested in #1 above, might be reasonable if the individuals purchasing service credit reflect the average plan member regarding life expectancy and salary progression, and if investment markets can be expected to provide returns matching the average long-term returns to the pension fund from the time of the service credit purchase until retirement. It is very unlikely that all of these assumptions will hold in a realistic situation. Commission staff has commented in the past that individuals will be more likely to purchase service credit if investment markets are depressed and are expected to stay depressed until the individual retires. The purchase price can be correct only if the pension fund can earn the plan’s assumed rate of return, generally 8.5 percent, on the assets received through receipt of the purchase price. If the investment markets are depressed and neither the individual nor the pension fund can earn that required return, then the purchase price is too low and the purchase is subsidized. Individuals with a longer than expected lifespan will be subsidized under a purchase because the payment of retirement benefits will continue for more years than predicated. Subsidies will also occur if the individual does not have a typical salary progression after the purchase occurs. Individuals expecting to soon receive a considerable promotion would want to purchase service credit before or soon after that promotion. The individual will ultimately have a high-five average salary considerably in excess of that predicted from the five-percent-increase assumption used in the pricing formula, creating a windfall.

   To guard against these effects, the Commission may choose to recommend using a rate of return assumption which is below the long-term expected return of the fund, mortality assumptions which predict a longer-than expected life, and a salary increase assumption that is above that used for the plan in general in the procedure to determine the purchase price. This would build in protection for the plan. Some individuals may wish to purchase service credit under these terms because of a strong preference for certainty. It is quite possible that, under these terms, individuals may find purchasing an annuity from the private sector has more value.

   Given the many general law service credit purchase provisions enacted in recent years, the public pension funds have, in effect, become general annuity providers, competing with private sector insurance firms and other private sector providers of annuities and investment products. The public pension funds are selling supplemental annuities to a limited group – only individuals who happen to be public employees – but the supplemental annuities are being justified based on service, or breaks in service, which have nothing to do with Minnesota public employment. The supplemental annuities are somewhat disguised because these annuities are rolled into and become part of the primary pension the individual earns for providing state or local service. In an effort to be sure the employee is not overcharged for the supplemental annuity, considerable effort is spent trying to refine the system used to price these annuities. In trying to be so precise, we risk harming the pension fund because there will always be flaws or weaknesses in the methodology. A private sector provider would recognize that it must build in a cushion to protect the company from market fluctuations, possible demographic characteristics of the covered group that differ from the general population, and increases in life expectancy which cannot be fully foreseen by the system used to price the annuity when that annuity is purchased. The suggestion is that if Minnesota public pension funds are to remain in the business of selling annuities to its members, an adequate cushion to minimize the chance of harming the fund should be built in.

4. Restrict Service Credit Purchases. Because service credit purchases may harm the pension fund due to flaws in the procedure used to compute the price, the Commission may wish to restrict the use of service credit purchases. The fewer service credit purchases that occur, the less chance there is that the purchases will create significant cumulative harm. These restrictions could be considered alone, or in conjunction with revised formulas to compute the service credit purchase prices. Some possibilities are as follows:

   1. Prohibit Purchases of Fractional Years of Service Credit. To date, most service credit purchases are for fractional years of service. Perhaps individuals are deciding to purchase just the service credit needed to qualify for early retirement and no more, because of the cost. Requiring purchases to be in full increments could decrease the number of purchases. Continuing to allow fractional years of service may also hamper any future effort to review the impact of service credit purchases, assuming budgetary funds become available at some point to pay for that effort. The actuary has indicated that the actuarial valuation is based on full year increments, and that a distorted comparison results if a service credit purchase price for fractional years of service is compared to the liabilities reflected in the next actuarial valuation related to that purchase. However, the second consideration memo noted that when fractional year purchases were eliminated from the existing data, typical differences between the purchase prices and the corresponding actuarial liabilities were as large or larger than before.

   2. Prohibit Purchases Just Prior to Retirement. The review in the second consideration memo indicated that many individuals who purchased service credit and then retired prior to the next actuarial valuation received large windfalls (the purchase price was much less than the computed resulting liability). To avoid harm to the fund, the Commission may wish to consider prohibiting purchases within one year, or possibly longer, before retirement. If retirement does occur during the specified period, the service credit obtained through the purchase could be forfeited with a refund of the purchase price.

   3. Consider Requiring Additional Contributions to Correct for Events After the Service Credit Purchase Occurs. The Commission may wish to consider requiring additional contributions to be made by those who purchase service credit to correct, to the extent feasible, for events occurring after the purchase and prior to retirement. Commission staff has noted in other documents that if a benefit improvement occurs after the purchase and the improvement does apply to the years being purchased, the individual receives a windfall. The Commission may wish to consider requiring a contribution from the purchaser to cover that windfall, or a requirement that the benefit improvement does not apply to purchased years of service credit. Perhaps seeking additional contributions from those with high-five average salaries higher than predicted might also be explored.

   4. Consider Permitting Service Credit Purchases Only as a Tool of the Employer to Address Budget Problems, Where that Employer Pays a Considerable Share of the Full Actuarial Value. Some contend that the numerous full actuarial value service credit purchase provisions are useful to create a more flexible teacher workforce to meet the employer’s needs. The argument is that individuals can retire after purchasing additional service credit and then return to provide service to their prior employer on a part-time or long-call substitute basis, helping the employer to provide needed services while helping the budget situation. An alternative view is that service credit purchases harm the employer, by causing teachers with long tenure and valuable skills to leave early and not return. If employers felt that the service credit purchase provisions provide the employer with a valuable workforce tool, one would expect many employers would be assisting employees to purchase service credit. The employer does have the authority to assist the employee with that cost. Both the 1998 method (Section 356.55) and the pre-1998 method (Section 356.551) currently permit an employer to pay the remainder of the full actuarial value purchase price, after the employee contributes the employee contributions that would have occurred if the period had been covered, plus interest. But the data suggests that there might be no cases to date where the employer voluntarily chose to help the employee cover the cost of a service credit purchase. Employers have contributed in a few cases, but there are so few that these might be due solely to situations where the Commission mandated an employer payment because the employer harmed the employee and admitted to that harm.

      Based on the study in the second consideration memo, and consideration of factors that cannot be captured in that data (such as investment returns and mortality that differ from the assumptions), subsidies may occur despite efforts to accurately estimate the full actuarial value. Thus, even with a full actuarial value approach, cost may be shifted from the employee to the pension fund and to all other contributors to the fund. Despite further efforts to refine the method, some potential for subsidy is likely to remain as long as service credit purchases continue. Therefore, the Commission may wish to consider, as an alternative to the largely unrestricted service credit purchase provisions enacted into law in recent years, restricting most service credit purchases to situations where the employer is willing to share in that cost. The employer should be willing to do so if the benefit outweighs the cost to the employer. This does not eliminate the risk of harm to the pension fund, but it does support a public purpose to offset the additional risk. Currently, individual employees are making service credit purchase decisions without consideration to the impact on the employer or on the pension fund.

   5. Further Restrict the Number of Years that May be Purchased. Many of the full actuarial value service credit purchase provisions in general law permit purchasing up to ten years of service credit. If subsidies can occur when an individual purchases one year or a fraction of one year, the subsidy is likely to be considerably greater the more years of service being purchased. If the general law service credit purchased provisions are to be continued in any form, the Commission may wish to consider reducing the number of years of service which may be purchased.

   6. Allow All or Several of the General Law Service Credit Purchase Provisions to Sunset. To the extent that service credit purchase provisions may create uncovered liability in the pension plan, leading to risk and financial burden for all employers and for the more than 99 percent of the plan membership who never use these provisions, and create inconsistent treatment between individuals who do use these provisions, the Commission may wish to consider allowing all or most of the general law provisions to sunset.

      If the Commission concludes that some should remain, the Commission may choose to restrict or revise those that remain. One source of guidance is the Commission’s Principles of Pension Policy statement. The statutory provisions that most strongly conflict with that policy statement could be allowed to expire. An earlier memo summarized each of these provisions and noted conflicts with the policy statement. The policy statement indicates that a period to be purchased should be previously uncovered public employment or quasi-public employment, and should have a Minnesota connection. Service credit purchase provisions which permit purchases of service credit for periods of private school teaching, or for various teaching or other service in other states or countries, clearly conflict with this policy. The provisions permitting full actuarial value purchases of service credit for periods of military service is another that creates policy conflicts. Individuals are eligible to use that provision to receive coverage for military service which was provided before the individual became a Minnesota public employee, and quite possibly before becoming a Minnesota resident. Those cases lack any Minnesota connection. In other cases, the individual was a Minnesota public employee before and after the military service, but the individual failed to take advantage in a timely manner (generally within five years of returning from the military service) of other provisions of the plan which permit the employee to receive service credit on very favorable terms. If an individual in this category had applied for special law authority to purchase service credit given that previous missed opportunity, the Commission might conclude that the special law request, following the failure to take advantage of an earlier opportunity to receive service credit at attractive terms, is a violation of equity considerations. The same consideration applies to maternity, paternity, or family breaks-in-service in cases where the lack of service credit stems from a failure to utilize in a timely manner other statutory provisions to receive the service credit.

There are several full actuarial value military service credit purchase provisions now in statutes, all set to expire in 2004. Provisions covering TRA and the first class city teacher plans were enacted in 1999, coded as Section 354.533 and 354A.097, respectively. Comparable provisions were enacted soon thereafter for the General State Employees Retirement Plan of the Minnesota State Retirement System (Section 352.275), the State Patrol Retirement Plan (Section 352B.01, Subdivision 3a), and the Public Employees Retirement Association (Section 353.01, Subdivision 16a). Copies of all of these provisions were attached to the first consideration memo.

These provisions are all similar. They permit vested members of the applicable plan to purchase service credit at the price determined under the 1998 full actuarial value method for "the initial period of enlistment, induction, or call to active duty without voluntary extension," providing the individual "is not entitled to a military pension" due to that service and the individual has "not purchased service credit from any other defined benefit public employee pension plan for the same period." The provisions therefore permit all initial military service periods to be covered, not for any period beyond that which the individual voluntarily agreed to serve. The language also attempts to prohibit double coverage by prohibiting any purchase if the service is covered by a military pension, or if coverage has already been established in some other plan.

TRA and the first class city teacher plans recently proposed changes in their provisions, and if the Commission were to conclude that changes are necessary for federal Internal Revenue Code compliance, comparable changes would also seem necessary in the similar MSRS-General, State Patrol Retirement Plan, and PERA provisions. The TRA and first class city teacher plan suggested changes were included in H.F. 1086 (Smith); S.F. 806 (Betzold): Various Plans; Uniformed Service Employment and Reemployment Rights Act (USERRA) and Internal Revenue Code Compliance, which was introduced during the 2003 Legislative Session and was among those bills referred to the Administrative Legislation Subcommittee. The Subcommittee took no action on these specific full actuarial value military service credit proposals because the changes have no effect if the provisions are not extended. Any change would presumably be effective in July 2004, which is after these provisions will expire. Therefore, the Subcommittee left this matter to be addressed by the full Commission as the Commission decides what action to take on the full set of full actuarial value service credit provisions, including those not related to military service, and the pricing provision.

The specific changes proposed by the teacher plans are to permit the purchase period to include periods of voluntary extension, to permit purchase of up to ten years of service credit, and to permit purchases for periods covered by a military pension. A copy of the proposal as it appeared in the Subcommittee’s first working document, LCPR03-220, is attached.

Whether the Commission will need to take any action on these military service full actuarial value provisions will depend upon what action the Commission decides to take on the full set of military and non-military full actuarial value service credit provisions. If the Commission decides to let all of the provisions expire, no action is needed. If the Commission does decide to keep some of these provisions, perhaps in a modified form, then some action may be needed at some point on these military service credit provisions. If some provisions permitting up to ten years of service credit are retained, the changes proposed for the military provisions may be appropriate if the Commission concludes that the discrimination concern raised by the pension fund administrators, and discussed below, has merit. If the Commission were to decide, however, to retain only a few of the provisions and to restrict the number of years that can be purchased to well under ten years, consistent with a suggestion made earlier in this memo, then this "ten year" military proposal proposed by TRA and the first class city teacher fund associations will need to be altered, assuming these provisions are to be continued in any form.

If the Commission decides to address these military service credit proposals, the first issue is whether USERRA or any other federal law requires these provisions to be revised as proposed. Commission staff is aware of no specific federal mandate requiring that a full actuarial value service credit purchase provision be included in a plan’s benefit provisions, and staff knows of no law specifically requiring that that a full actuarial value military service purchase provision must permit purchasing up to ten years of service credit and include periods of voluntary extension.

To the best of staff’s knowledge, the proposed change to permit up to ten years of service credit purchase and to include periods of voluntary extension are based on a general concern expressed by TRA and the St. Paul Teachers Retirement Fund Administration that ex-military personnel should not face discrimination. There are other full actuarial value service credit purchase provisions in TRA and first class city teacher law (the private or parochial school teaching provision, for example) which permit a purchase of up to ten years of service. The current law version of these full actuarial value military service credit purchase provisions allow a purchase for only the initial period of military service without any voluntary extension of service. This is almost always less than ten years. Therefore, the argument is that to avoid a discrimination claim, the military service full actuarial value service credit purchase provision should be revised to allow up to ten years to be purchased, consistent with the most generous of the other non-military service credit purchase provisions. Since that military service period will almost never last for ten years unless a voluntary extension is involved, the proposal also struck the requirement that the period could not include a voluntary extension.

The prohibition against purchases if the period is covered by a military pension, which was an effort to avoid any appearance of double-dipping, is proposed to be stricken based on concern that the courts might conclude that the current prohibition conflicts with federal law. A brief memo from Jon Murphy, Assistant Attorney General, to Luther Thompson, TRA Assistant Executive Director, dealing with this matter is attached.

This memo notes the current situation regarding the 1998 full actuarial value pricing procedure and the various general law service credit purchase provisions enacted in 1999 and later. Given that all of these provisions are set to expire in May or June of 2004, the memo describes various options for the Commission’s consideration. If the Commission concludes that the currently used full actuarial value estimation method and the various service credit purchase provisions added to statutes during recent years should be permitted to expire, no legislation is necessary. If the Commission wishes to pursue some other course of action, the Commission can provide specific direction to staff to draft applicable bill language.

Attachments: