At the December 11, 2001, Legislative Commission on Pensions and Retirement (LCPR) meeting the Commission reviewed the first consideration memo dealing with the ongoing funding issue of the General Employee Retirement Plan of the Public Employees Retirement Association (PERA-General). That memo reviewed the pattern of PERA-General contribution level funding deficiencies/sufficiencies from 1986 through the preliminary July 1, 2001, results. Another section discussed the 1996 and 2000 PERA-General experience study results and the difficulties in effectively measuring PERA-General member turnover. That memo also noted some issues with the new actuarial value of asset methodology, first used for the July 1, 2000, actuarial reports. The change from the prior actuarial value method added $44.9 million to the PERA-General computed unfunded liability. LCPR staff also alerted the Commission of the need to be aware of the market value of various Minnesota public pension plans in addition to results based on the actuarial value. The actuarial value methodology is an attempt to smooth results, but it does not produce a conservative estimate of total assets during weak market periods. Given the severe downturn in the equity markets in calendar years 2000 and 2001, the actuarial value of assets is greater than the market value of assets. That leads to higher computed funding ratios and less unfunded liability within the pension systems than would be computed given the actual assets of the pension funds on the valuation date (July 1, 2001).

The current memo is for the Commission’s second consideration of PERA-General funding issues. In this memo, staff compares the preliminary July 1, 2001, PERA-General actuarial report results (the results of the draft report that was available when the first consideration memo was written), with the results of the final July 1, 2001, PERA-General actuarial report. Results from earlier PERA-General actuarial valuations are also included to permit the Commission to note patterns over time. Finally, we compare the most recent actuarial valuation report results, which follow from the actuarial value of asset methodology, with results based on market value. This information is provided not only for PERA-General, but also for all the Minnesota public pension general employee plans (General State Employees Retirement Plan of the Minnesota State Retirement System (MSRS-General), PERA-General, the Teachers Retirement Association (TRA), the first class city teacher plans, and the Minneapolis Employees Retirement Fund (MERF)). We currently do not know whether there will be a quick recovery from the market downturn that occurred in calendar years 2000 and 2001. Given this uncertainty, the Commission may wish to consider two sets of results, one set based on actuarial value and another based on current (as of July 1, 2001) market value.

PERA-General’s actuarial condition, as reported in LCPR staff’s first consideration memo based on the preliminary July 1, 2001, actuarial report, is noted in the first column of Table 1. The results in the second column are from the final report. LCPR staff received that final report a few days before the last LCPR meeting, but after the much of the first consideration memo was written. Some information from the final report was included in a attachment to the earlier memo, and the Commission heard brief testimony from Ms. Mary Vanek, PERA’s Executive Director, and from PERA’s actuary, Ms. Bonnie Wurst from the William M. Mercer Company, regarding member data issues.

It was evident from member count comparisons between the preliminary and final July 1, 2001, actuarial value reports that PERA’s member data are problematic. This was reinforced by testimony provided by Ms. Vanek and Ms. Wurst at the last Commission meeting. The testimony indicated that PERA is having difficulty determining which individuals have terminated PERA-covered employment and need to be moved to the deferred annuitant or non-vested terminated status. In other words, PERA is having difficulty accurately determining its turnover. This is important because the termination (turnover) assumption used by the actuary has a significant impact on PERA-General’s computed funding requirements. Revised turnover assumptions are part of changes that created a large computed contribution deficiency (1.96 percent of payroll) in the July 1, 2000, PERA-General actuarial valuation. We know from recent experience studies and related correspondence that the actuary retained by the LCPR has had difficulty identifying the most reasonable turnover assumptions to apply to this plan. The 1996 experience study indicated less turnover than previously assumed, and revised turnover assumptions were adopted and used in the July 1, 2000, actuarial valuation. A recently completed experience study for 2000 is now indicating more turnover than expected. The difficulty in accurately predicting PERA turnover may reflect problems with the turnover data used in the experience studies. If the data are inaccurate, it will not be possible to compute the actual turnover that is occurring, nor to be certain of any trends suggested by the data. What appears to be a change in termination behavior may reflect nothing more than an improvement (or worsening) of data quality.

Regarding the PERA member data quality issue, in testimony Ms. Vanek indicated that PERA has problems with member reporting by school districts. These employers may have cafeteria employees, school bus drivers, and similar employees on stand-by or on-call status. The individuals work when needed, but there may be periods months, or more than a year, where the individuals do not work for the public employer and thus receive no salary from the public employer. This presents a challenge for handling these employees within an actuarial valuation.

Ms. Wurst indicated in her testimony that the active employee total in the PERA July 1, 2001, preliminary actuarial valuation report produced by Milliman USA was overstated. For the final report, the active member file was again reviewed and any individual for whom there was zero salary for the year and zero service credit was shifted to one of the terminated categories. If the applicable individual had sufficient past service to vest, she was moved to deferred retiree status, and if the individual was not vested, she was shifted to the non-vested former member category. While this action by the actuaries is not unreasonable, it will create an overstatement of deferred and non-vested former members and a corresponding understatement of active members. In all these cases PERA has no documentation from the employer indicating that these individuals have actually terminated service. Some of these individuals are stand-by or on-call status employees that Ms. Vanek mentioned in her testimony, who provided no service during the year under study. These individuals will pop back into the active member category when they are called by the employer and salary payments occur. Ms. Wurst also mentioned situations were the records for individuals indicate that service credit was earned or granted but for some unspecified reason no salary data appeared. Ms. Wurst indicated that in those cases, the individual presumably was retained in the active member category and the actuary took salary earnings in a prior year and used it as a proxy for the unknown current year salary.

The various agreements between the PERA-retained actuary and the LCPR-retained actuary just described are reflected in the final July 1, 2001, PERA-General actuarial valuation information shown below and influence the differences between the preliminary and final reports. Regarding the specific differences between the two reports as shown in the following table, the preliminary report indicated 151,413 active members while the final report lists only 138,759, an adjustment of 12,654. That adjustment reflects moving more than 3,500 individuals to deferred retiree status, and moving more than 8,300 people to the non-vested former member category. That left an additional 728 individuals that were removed from the active member total but not added to any of the terminated employee categories. Ms. Wurst testified at the December meeting that these were duplicate files and were eliminated.

The change in covered payroll between the preliminary and final reports is surprising in its size. The covered payroll figure in the final report is $15.4 million higher than in the preliminary report. Ms. Wurst mentioned situations where the prior year salary was used for individuals where PERA records indicated some service credit during the year under study but no salary. Whether that is the only cause of this adjustment or whether there are multiple causes, the combined effect is significant. The Commission may wish to seek some further comments from PERA’s Executive Director or from the PERA-retained actuary or Commission-retained actuary regarding the cause or causes for the increase in covered payroll between the preliminary and final reports.

Despite the extensive revisions in the membership categories and the and covered payroll between the preliminary and final report, the computed change in the expected contributions and the contribution requirements were minimal and were offsetting, leading to no change in the estimated contribution deficiency, 1.28 percent of payroll.

Table 1

PERA-General
July 1, 2001, Actuarial Valuation Results

The following table displays the results of the PERA-General actuarial valuation reports from 1994 through the recently received final version of the July 1, 2001, report. The 1994 through 1997 reports indicate contribution deficiencies although the pattern hinted at an improving trend. The contribution deficiency in 1997 was 0.33 percent of payroll, a lesser deficiency than the prior three reports. The 1998 report suggested that the situation had improved to a modest contribution sufficiency of 0.19 percent of payroll, or $6.6 million. The contribution sufficiency increased to 0.57 percent of payroll in the 1999 report. The situation reversed in the July 1, 2000, actuarial valuation. While the 1999 report indicated that contributions were $19.7 million (0.57 percent of payroll) more than needed to fully fund PERA by the full funding date (which at that time was the year 2020), the July 1, 2000, actuarial valuation suggested that contributions were 1.96 percent of payroll or $70.4 million too low.

The considerable difference between the 1999 and 2000 actuarial valuation reports was influenced by revised actuarial assumptions for PERA-General, which were first used in the 2000 PERA-General actuarial valuation, and the use of the newly revised actuarial value of assets procedure. Those changes contributed to considerably higher estimates of PERA-General unfunded liability (the computed unfunded liability was about $570 million higher than in the prior year).

In reviewing the PERA-General 2000 actuarial results in the table, Commission members will also note that the amortization date was extended to 2024, from 2020 in the prior valuation. That extension reflects the operation of Minnesota Statutes, Section 356.218, Subdivision 4g, paragraph (c), which automatically extends the amortization period whenever revised actuarial assumptions, a revised actuarial cost method, or revised benefit plan provisions creates an increase in unfunded liability.

The final 2001 PERA-General actuarial valuation suggests a contribution deficiency which is not as large as that computed for the 2000 report. The 2001 contribution deficiency officially is 1.28 percent of payroll, or $49.2 million. These 2001 results reflect part of changes enacted during the 2001 First Special Session. Legislation passed during that special session increased the PERA-General employee and employer contributions by 0.35 percent of salary (when combined totaling 0.70 percent of salary). These higher rates were effective on January 1, 2002, which is in the middle of PERA’s fiscal year. The actuary handled this situation by assuming in the report that employee and employer contributions reflect only half of the increase. The report notes in a footnote that if the contribution rate increase is assumed to be in effect for the whole period (from July 1 rather than from January 1), the contribution deficiency would be about 0.95 rather than 1.28 percent of payroll. Looking into the future, 0.95 percent may be the better estimate. The amortization date in the 2001 actuarial report is December 31, 2031. This revised date is caused by action by the 2001 Legislature, which included a provision specifically extending the amortization date to 2031 as part of the PERA-General remedial provisions passed during the 2001 Special Session.

Ms. Vanek indicated at the Commission’s recent December meeting that PERA does not intend to seek any further increases in PERA contribution rates or state aid during the coming legislative session. This may reflect a decision on PERA’s part to not seek further action until a full year of experience at the new, higher contribution rates has occurred, and pending any possible additional actuarial assumption changes. The recent 2000 PERA experience study results suggests that new actuarial assumption recommendations may be forthcoming from the actuary. The decision may also be influenced by the budget condition of state and local governments.

Table 2

PERA-General Actuarial Condition
1994-2001

As previously mentioned, at the current time the actuarial value of assets for Minnesota public pension funds currently tends to be greater than the market value of assets due to the recent negative returns earned in equity markets. If the economic and market downturn is prolonged, sole reliance on actuarial value may delay recognition of a significant change in the funding condition of Minnesota’s public plans. In the tables in this section, we present results from the July 1, 2001, actuarial valuations for the general employee plans: MSRS-General, PERA-General, TRA, the first class city teacher plans, and MERF, and we contrast these results with results that follow from use of market value. For all these plans the market value was less than the actuarial value. The comparisons in this section serve not only to highlight the differences between the two approaches, but also to alert the Commission to funding issues in some other pension plans. This may be a consideration in any decisions the Commission and Legislature make in choosing how to allocate the state’s scarce financial resources. For plans with contribution sufficiencies, the sufficiency is more modest. Similarly, for any plan with a contribution deficiency, the contribution shortfall is larger than suggested by the use of actuarial value.

Consider a plan like PERA-General, which is less than fully funded. A pension fund’s unfunded liability is computed by subtracting current assets from the accrued liability. "Current assets" is defined in law as the actuarial value of assets. If we instead define current assets to be the market value of assets and the market value is less than the actuarial value, the computed unfunded liability will increase. Since there is more unfunded liability, the amortization contribution increases. The actuary would compute the required amortization contribution by dividing the computed unfunded liability by the present value of future payrolls through the amortization date.

MSRS-General is an example of a plan that has assets in excess of liabilities (in other words, it has negative unfunded liability). Under current law, the actuary would compute a negative amortization rate. Since a plan’s funding requirement is the sum a the normal cost, expenses, and amortization requirement, a negative amortization requirement acts to reduce the total required contributions to an amount that is less than the sum of the expenses and normal cost. For purposes of the actuarial work, the amortization rate is the negative unfunded liability (the assets in excess of liabilities) divided by present value of future payrolls assuming a 30-year amortization period.

Results for PERA-General appear in Table 3. If PERA-General’s market value were used, the plan’s unfunded liability would increase by $436 million and its funding ratio would fall, the amortization requirement would increase by 0.54 percent of payroll, and the contribution deficiency would be 1.82 percent of payroll rather than 1.28 percent of payroll.

Table 3

PERA-General
July 1, 2001, Actuarial Valuation Results
Compared to Market Value Results

Results for MSRS-General, shown in Table 4, indicate that use of market value lowers the plan’s funding ratio from 112.07 percent funded to 106.05 percent. The plan continues to have assets in excess of its liabilities, but that excess has decreased noticeably. Rather than having negative amortization equal to 2.17 percent of payroll to offset the plan’s normal cost and expenses, the negative amortization rate is reduced to 1.09 percent of payroll, leading to a higher total contribution requirement. The plan would have a modest contribution sufficiency of 0.13 percent of payroll, rather than a contribution sufficiency of 1.12 percent of payroll.

Table 4

MSRS-General
July 1, 2001, Actuarial Valuation Results
Compared to Market Value Results