A. Scope

W.S. 28-8-107 (b) authorizes the Legislative Service Office to conduct program evaluations and performance audits. Generally, the purpose of a program evaluation is to provide a base of knowledge from which policy makers can make informed decisions.

In June 1996, the Management Audit Committee selected the topic of cost-of-living adjustments under the Wyoming Retirement System as a subject for review. This system encompasses four separate retirement programs: the Public Employees' Pension Plan; the Paid Firemen's Pension Plan; the State Highway Patrol, Game and Fish Warden, and Criminal Investigator Retirement Plan; and the Volunteer Firemen's Pension Plan. We limited the scope of our work to the Public Employees' Pension Plan, because it is the by far the largest retirement program and has been the subject of the most legislative attention. Accordingly, the research described in this report centered around the following questions:

• Specifically, what do system administrators and policy makers hope to accomplish by providing post-retirement pension increases under the retirement system?

• What have been the cumulative financial effects of these post-retirement pension increases on individual retirees and the system?

• Do responsible officials have access to sufficient management information to support effective decision making?

• Does the state have appropriate mechanisms to coordinate the actions of administrators, the board, and the Legislature?

• What are the advantages and limitations to Wyoming's approach to funding cost-of-living increases?

• Are there alternative funding strategies that should be considered?

B. Methodology

This evaluation was conducted according to statutory requirements and professional standards and methods for governmental audits. We conducted our research between June and August 1996.

We reviewed relevant statutes, rules, regulations, policies, annual reports, budget documents, professional literature, board minutes, annual financial reports, other reports and studies, and a variety of other program materials. In addition, we reviewed information extracted from the system's computerized database and original records for a sample of 125 randomly-selected retirees.

We attended a board meeting and interviewed several current and former board members on an individual basis. We also interviewed key system administrators and various other state officials with knowledge of the retirement system. Finally, we conducted surveys of Wyoming pensioners and retirement system officials in other states.

C. Acknowledgments

The Legislative Service Office would like to express appreciation to the individuals who assisted in our research, especially to the staff and board of the Wyoming Retirement System.

To mitigate the effects of inflation on the pensions of retired public employees, many retirement systems provide periodic cost-of-living adjustments (COLAs). Without increases, even at today's moderate rate of inflation, the value of a pension after several years of retirement could be far less than its value at the time of retirement. Such economic security protection is an issue for pensioners and policy makers, considering that the period of retirement often extends for some 20 years or more. Figure 1 shows the percent annual change in the rate of inflation for the ten-year period from 1986 through 1995, as measured by the Consumer Price Index (CPI).

There are two basic approaches used to provide cost-of-living adjustments, ad hoc and automatic. Ad hoc increases are made at irregular intervals and employ various ways of determining the amount of increase provided. Frequently, ad hoc adjustments involve the use of flat-rate formulas, fixed or graded percentages, or a combination of the two. An ad hoc COLA is discretionary, may not be awarded to everyone, and is not built into the financing mechanism of a retirement system. Many employers find ad hoc COLAs attractive because of their controlled, predictable cost and the fact that costs will decline over time, as the beneficiary groups grow smaller.

An automatic COLA provision in a pension plan means that benefits for all current and future retirees are automatically adjusted, usually according to an increase in the CPI. An automatic COLA is built into the plan's financing mechanism. Pension benefits that are automatically indexed to inflation are ideal from a retiree's perspective. However, because the costs can be staggering, full indexing is rarely provided. An annual compounded increase of three percent in a pension, for example, would increase the long-term cost of a plan by 30 percent. Thus, to assure the financial health of a retirement system in the face of an unknown future rate of inflation, a system may place a cap on the maximum increase. Systems may also decide to grant automatic COLAs only if the CPI increases by a certain minimum amount, say three percent. According to professional literature, the basic philosophy of plans that put caps or limits on COLAs seems to be that an employer and employee should share the risk of inflation.

Figure 2 shows how the purchasing power of fixed retirement income declines under different rates of inflation. For example, at a steady rate of three percent inflation, the purchasing power of initial benefits would be reduced by 25 percent in 10 years.

The Social Security system and pension plans for federal workers incorporate automatic, annual COLAs. Further, over half the states reporting information to the Bureau of Labor Statistics in 1992 indicated they also provide automatic COLAs annually. These COLAs were generally capped at between three and five percent. Most of the remaining states reported they traditionally provided ad hoc COLAs.1 However, because of the lower inflation since the 1980s, the number of states granting ad hoc COLAs has gradually decreased since 1987.

Wyoming Retirement System

Nearly all public employees in the state, excluding federal workers, are eligible to participate in the Wyoming Retirement System (WRS).2 This includes full-time and regular part-time employees of the state, public school districts, cities and towns, and institutions of higher learning. At the end of 1995, there were 432 Wyoming government employers participating in the system. School system employees are the largest group within the system, making up 48 percent of the active membership. State employees comprise 24 percent, with other groups making up an additional 28 percent of membership.

As of December 31, 1995 there were 31,804 employees in active service covered under WRS. Of these employees, 1,023 were law enforcement officers who are eligible for special benefits under provisions of the plan. There were also 3,562 inactive, non-retired, vested members who did not elect to receive their accumulated contributions when they left covered employment. In addition to these active and inactive workers, a total of 11,864 pensioners and beneficiaries were receiving monthly payments.

Basic Benefits. WRS is a "defined benefit" system in which eligible employees earn retirement benefits based on a formula specified in statutes. Retirees are owed these benefits as a matter of law. Under this type of system, benefits are not directly related to employer and employee contributions. For each year of credited service, members earn benefits equal to two percent of their highest average salary (for a continuous 36-month period). Employees are "vested" in the system, meaning they earn the right to benefits at retirement age, after 48 months of service. However, to begin drawing a monthly benefit check, members must be age 50 with at least four years of service, or at any age with at least 25 years of service. Normal retirement occurs at age 60, although law enforcement officers receive unreduced benefits at age 50 with 25 years of service. Under board rules, benefits are reduced by five percent per year for retirement prior to that age. For example, an employee whose highest average salary was $20,000 retiring at age 60 after 20 years would receive a pension equal to 40 percent, for a monthly benefit of $667. Members can also retire with full benefits if they satisfy the "rule of 85," which requires years of service plus age to total 85.

Prior Benefit Structures. WRS began as a teacher retirement system in 1943. Under this original system, members contributed one percent of their salaries and earned a pension equal to $2.50 per month for every year of service. In 1949, state employees entered the program. Members became eligible for Social Security in 1953 and the system was converted to a "money purchase" plan using the amount of contributions paid and age at retirement as a basis for benefits. Although the current benefit formula was added in 1975, members with service under the old plan were provided options for how their pensions would be calculated to produce the largest monthly benefit.

Retirement System Administration

Wyoming Retirement Board. The Wyoming Retirement System is guided by an 11-member board, which includes the State Treasurer, two public employees, two employees from the public school system or the University of Wyoming, one retired member of the WRS, and five others who are not members of the retirement system. The board, with the exception of the State Treasurer, is appointed by the Governor with the consent of the Senate. Board members serve six-year terms.

The board meets bi-monthly, and is responsible for the control and management of the system's assets. According to the WRS strategic plan, the role of the board and staff is to safeguard the financial integrity of the system through prudent management, and provide adequate member benefits within the limitations of system funding.

WRS Staffing. The board employs a director, who serves as the board's secretary. The director is responsible for the day-to-day operation of WRS, its 17 full-time employees, and contractors. The office has four sections: administration, active employee, benefits and payroll, and finance and accounting. Total administrative expenses for all four systems, excluding contract services, were $780,531 during 1995 and were paid from system assets, not the General Fund.

Contract Services. In addition to its administrative staff, the board employs a number of contractors. These professionals include a consulting actuary, who determines the amount of money that will be necessary to pay pension obligations. The actuary makes assumptions about investment returns, salary growth, employee turnover, life expectancy, and other factors, to determine the system's financial liabilities. The actuary also determines the cost of proposed benefit enhancements.

The board employs an investment consultant to oversee the performance of the system's eight contracted investment managers. The eight managers are responsible for day-to-day investment decisions concerning the system's assets. The board also contracts for data processing, an annual financial audit, and other services to administer the four systems. In 1995, the total cost for these contractors, excluding the investment managers, was $261,260, while fees to the investment mangers totaled $4,676,943.

Retirement System Funding

The funding of WRS is based on the principle that pension funding should be a contemporary obligation. This requires setting aside adequate funds to offset accruing pension liabilities for today's workers. The advance funding principle of WRS is the opposite of "pay-as-you-go," where all funding begins after an employee's service is completed and a pension is earned. Although the latter process is widely used by the federal government, it has the disadvantage of deferring obligations to future generations of taxpayers.

Although few states operate entirely on a pay-as-you-go basis, many have used this approach in part. This occurs either by regularly appropriating less than the full cost of the benefits, or by occasionally underfunding this cost. Underfunding public pension plans can create severe future financial difficulties for retirement systems.

As specified in W.S. 9-3-412 and 413, WRS is supported by contributions equal to 11.25 percent of the total payroll for all eligible employees. Of this statutory contribution rate, 5.68 percent is considered the employer's share and 5.57 percent is the employee's share. However, the employee's share is commonly "picked-up" or paid by most public employers in the state. During 1995, total contributions to the four systems amounted to just over $96 million. In addition, WRS invests those assets which are not required to pay benefits to current retirees. Total return to the system from investments was over $578 million during 1995.

System investment returns are important since WRS is a "mature system," meaning that in each year more money goes out in benefit payments than comes in through active member contributions. Hence, the system will increasingly rely upon investment earnings to fund benefits.

Figure 3 provides a summary of the financial condition of WRS for the last several years. It shows the system's unfunded liability (which was the result of pension liabilities accruing faster than assets) decreasing and the funding ratio (assets as a percent of liabilities) improving in recent years. While there were some fluctuations due to investment earnings, the overall trend has been towards a stronger financial position.

System Post-Retirement Adjustments

The Legislature granted numerous post-retirement pension increases, which are commonly referred to as COLAs, between 1978 and 1994. Session laws for those 17 years indicate that eight ad hoc zincreases and one automatic COLA were approved for retirees under the main system.3 Only one of these bills, from 1981, had committee sponsorship and resulted from a legislative interim committee study; the rest were individually sponsored or co-sponsored.

The Legislature passed another 38 bills during this period, making various other changes to the system; many of these bills, directly or indirectly, had an effect on COLAs. For example, they improved survivor benefits, expanded membership, enacted residency requirements, required use of a unisex mortality table, offered an early retirement option, and charged the board with annually determining any change in the year's cost-of-living. In addition, a 1990 amendment to the Wyoming Constitution requires retirement funds be used only for the benefit of members, retirees, and beneficiaries of the system.

Collectively, this history reflects the Legislature's ongoing interest in COLAs and its willingness to approve adjustments. The most recent COLA was approved in 1994, granting those who retired at various intervals before 1980 an additional $2 per month increase for each year of service.

Ad Hoc Adjustments. The Legislature has used several different criteria to target ad hoc adjustments: specific retirement dates, service between certain dates, the number of years of full-time service, and whether the service was provided before 1975 (when the current benefit structure was adopted). Altogether, the Legislature created 18 sub-groups of retirees through ad hoc adjustment legislation. As a result, benefits for many retirees were increased by more than one ad hoc adjustment. For example, one ad hoc adjustment targets 1975-1980 retirees with service prior to 1975; another benefits persons retiring before 1984, and yet others benefit all pre-1980 retirees. Currently, those who retired as long ago as 1953 and as recently as 1990 are receiving ad hoc adjustments.

The ad hoc adjustments have been made as both flat-rate and percentage increases. The most common adjustment has been a $2 increase to the monthly benefit, for each year of service. Less frequently, the Legislature has authorized percentage increases to adjust the benefits for specific groups of retirees.

Automatic COLA. In 1989, the Legislature passed W.S. 9-3-419(b) establishing the system's annual, automatic cost-of-living adjustment. This measure increases the benefits of system members who have been retired for at least two years by the lesser of the previous year's Wyoming cost-of-living index, or one percent, if the actuary determines it is actuarially sound to do so. Since this COLA became effective in 1991, an automatic one-percent increase has been granted to all eligible members annually.

COLA Costs. As illustrated in Figure 4, during the last ten years the annual cost to provide ad hoc adjustments has been far greater than the cost of the automatic COLA. However, WRS managers expect future costs to provide the automatic one-percent COLA to increase as more of today's workers retire and become eligible for this benefit. If no additional ad hocs are granted, the costs for these increases should decrease as the number of eligible retirees grows smaller.

The first-year costs of all ad hoc increases before 1991 were paid from the General Fund, with subsequent costs absorbed by the system. Since then, the retirement system has absorbed the full cost of all ad hoc increases for retirees. WRS managers reported the cost of ad hoc adjustments granted during the ten-year period ending in 1995 totaled approximately $60 million. Since the automatic COLA became effective in 1991, an additional $6.8 million of system funds have been expended providing one-percent compounded increases to all eligible members.

Figure 4 also shows the total cost for both ad hoc and automatic COLAs was $12.4 million in 1995. Of that, $9.8 million (79 percent) funded ad hocs and $2.5 million (21 percent) funded the automatic one-percent increase.

WRS Board Goal. The board has established a goal of increasing the cap on the automatic COLA from one to three percent when the system can afford to do so. Furthermore, the board is on record as opposing the provision of additional flat-rate ad hoc adjustments, believing that funding these increases thwarts the system's ability to afford an increase in the automatic COLA cap.

Improvements are Needed in the Approach to Provide COLAS

We examined policies and practices affecting COLAs provided under the main retirement program of WRS. We found the retirement system's financial condition has improved markedly in recent years. WRS should be in a strong position in the future to provide promised pension benefits to eligible public employees when they retire.

We also found that system assets have been used to support numerous cost-of-living adjustments for current retirees. Our analysis showed that these increases have been largely successful in mitigating the effects of inflation for many long-term retirees. However, our examination also identified several areas in which the Legislature and the WRS can work together to improve the state's approach to providing future COLAs.

First, and fundamentally, an overall policy framework is needed for determining the purpose of future cost-of-living adjustments. Also, we found the Legislature and WRS can improve their approach to providing future COLAs by:

• Enhancing management information to aid the decision making process.

• Developing a structured process to facilitate legislative consideration of COLAs.

• Identifying achievable ways to fund future benefit enhancements.

In the following sections of this report, we discuss these issues and make recommendations for improvements for the consideration of WRS staff, the board, and the Legislature.

Finding 1: Policy Makers Have Not Agreed on a Clear Purpose for COLAs

Between 1978 and 1994, the Legislature increased retiree benefits nearly every biennium. Further, it is likely that legislators will be pressed to continue this stream of legislation. Costs for providing these benefit improvements have increased to more than $12 million per year. However, without a relevant policy framework stating a purpose for these benefit improvements, there is no way to assess whether these expenditures are accomplishing the desired result for the desired beneficiaries.

Our analysis suggests that the legislation passed to adjust retirement benefits may have had different purposes. First, there may have been an intent to benefit those who worked and contributed under the money purchase plan. Second, providing long-term retirees with some insulation from the effects of inflation may have been a goal. Third, legislation may have been intended to improve the very small pensions of some retirees, regardless of which system they served under. While these purposes are not necessarily in conflict, they were not woven together to create an overarching policy framework to provide ongoing guidance to decision makers.

COLA Legislation Has Had Different Effects

Most of the ad hoc adjustments have benefited individuals who retired before 1980 and who would likely have worked most of their careers under the money purchase system. Because contributions were capped at specified salary levels (for example, at five percent of $10,000 in 1974), benefits resulting from this system were low. Retirement system officials told us that the Legislature has been very generous in compensating for these small benefits. Our sample analysis showed that adjustments make up 83 cents of each dollar of the current median benefit for those who retired before 1976, and 71 cents of each dollar for those who retired between 1976 and 1980.

By giving larger increases to groups of retirees who were retired longer, the Legislature may have been attempting to mitigate inflation's effect on purchasing power. Most ad hoc adjustments have done this, apparently to help long-time retirees "catch up" to a level of purchasing power that was not specified. The Legislature more clearly implied the intention of compensating for inflation in the legislation establishing the system's automatic COLA by indexing the increase to the Wyoming cost-of-living index, capped at one percent.

Flat-rate adjustments most benefit those retirees with small pensions. They provide a specified dollar amount per month for each year of covered service. Thus, employees retiring at the same time with the same length of service receive the same increase, regardless of the size of their initial pensions. By granting these types of ad hoc increases, the Legislature may have been compensating for small pensions, rather than addressing the purchasing power of all pensions.

In contrast to this historical legislative trend, the board opposes granting any more flat-rate ad hoc adjustments. The board envisioned and supported the automatic one-percent COLA as a mechanism to provide cost-of-living adjustments to all retirees, present and future. An automatic COLA, which brings all members up by the same percentage of their pensions, does not "catch-up" benefits for specific groups nor compensate for pensions that were small to begin with. The board believes ad hoc adjustments actually work against increasing the automatic COLA, since they divert necessary funding.

Important Questions Have Not Been Addressed. While legislation has accomplished a great deal of pension adjustment through ad hoc measures, there has been no encompassing plan or purpose guiding these measures. Instead, the legislation has raised different policy questions: Should the system proportionately improve all members' pensions through automatic percentage COLAs, or should it periodically "catch-up" long-time retirees? If so, to what extent? Should the system provide pensions of some minimum amount to members who retired long ago? In general, what is the purpose behind the pension adjustments? These policy questions are, as yet, unanswered.

Political Considerations Determine the Actions Taken

The Legislature has initiated most COLAs in response to constituent requests. Many of the individuals we interviewed sensed a great deal of retiree pressure on legislators to improve benefits. A representative of an employee group felt this pressure came as much from individual retirees as from organized retiree groups: "They go to their individual legislators with their stories and often, the members respond with a bill." WRS officials perceive that ad hoc adjustments have benefited "pockets" of retirees with the most clout, and that priorities for future enhancements will be set by those groups with the greatest voice.

Heightened Retiree Expectations

The lack of a policy framework may create heightened assumptions among retirees about what increases they will receive and how often these will occur. In our mail survey of 125 retirees, nearly half the respondents said they believe that state officials have indicated a commitment to maintain the full purchasing power of their pensions. Other retirees felt increases should occur more frequently than they have.

The specter of inflation eroding the purchasing power of retirement income is a constant concern of retirees. The widely known fact that Social Security benefits are indexed to the CPI, and the introduction of the Wyoming cost-of-living index as a basis for benefit adjustment, may lead both retirees and active members to expect that their pensions should be maintained, at least to an extent greater than the one percent specified in statute. In turn, these expectations lead to political pressures.

Further, without an indication of what can be expected from COLAs, employees may not take responsibility to provide for a portion of their retirement incomes while they are still working. A principle underlying defined benefit plans is that employees, at retirement, will receive lifetime benefits proportionate to their highest salaries and lengths of service. Some active members may assume that because they belong to a defined benefit plan, this proportionate benefit will be maintained throughout their retirement.

Policy Is Necessary For Fair and Effective Benefit Programs

Professional literature recommends that benefit plans have policies establishing their basic principles and fundamental purposes. Statements of benefit policies may range from expressions of philosophy to specific guidelines for plan management. However specific these may be, all professional sources we reviewed agree that an articulated policy subject to review is necessary to develop and maintain a fair and effective benefit program. Since COLAs are important parts of benefit programs, their implementation should also be guided by policy.

The National Conference of State Legislatures (NCSL) has a working group on state retirement systems. This group maintains that a general policy framework provides a basis for choosing among numerous and often conflicting proposals for change. In addition, policy statements clarify and communicate benefit management accountability, and they establish standards for measuring program performance.

Once established, policies tend to be relatively permanent, but they still require periodic review. It is important for decision makers to recognize circumstances that render some policies inappropriate or that create a need for new policies.

Wyoming statutes imply a basic retirement benefit policy by establishing a benefit formula and an automatic one-percent COLA. Furthermore, the Legislature's implementation of ad hoc benefit improvements in response to continuing pressure also implies a policy, albeit a patchwork approach. This policy has evolved by default through legislation that grants benefit improvements to specific groups of retirees who present their needs to the Legislature. It requires only that benefit improvements be advocated and affordable.

While acknowledging the importance of affordability as a criterion for setting policy, the NCSL working group concluded that:

...the actual cost of a provision should be considered secondary to plan design. The fact that a legislative proposal is not exorbitantly expensive is a weak argument for its adoption. Fairness and consistency with established principles of pension policy are most important.

There are no national standards for benefit formulas or COLA provisions. Nonetheless, some states have developed policy guidelines to help them address COLAs in ways that satisfy their particular needs and political philosophies. For example, some retirement systems have policies to target retirees whose benefits have been most deflated by inflation. These systems establish benefit "floors" to determine which retirees receive ad hoc adjustments. Along these lines, the Public Employee Retirement System of Idaho approved a minimum 80 percent restoration level for all members retired prior to 1980.

A retirement study commission in Ohio adopted a set of pension principles to assist its Legislature in effectively managing public pension plans. One principle states that:

• The benefit should be adequately maintained in retirement.

• A plan is needed to pre-fund post-retirement benefit increases.

• When possible, increases should follow a recognized economic indicator.

• Adjustments should primarily address the effects of inflation and not offset the effects service and salary had in determining initial benefits.

Maine has set a general objective of providing career employees with sufficient income in retirement to maintain their pre-retirement living standard. To meet this objective, the Maine State Retirement System has determined that an appropriate target for benefit adequacy is to have no reduction in the pre-retirement standard of living for an employee retiring at age 65 with 35 years of service and a reasonable amount of personal savings. The system has calculated that this target can be met if lower-paid employees have been able to save five percent of pay while working.

Neither the Legislature Nor the Board Has Defined the Outcomes Intended

Although both ad hoc and automatic COLAs have been granted, there has been no comprehensive effort to describe the needs or problems they are intended to solve. The first step in developing a policy is identifying the problem and what is causing it. With that done, policy makers can review the results or outcomes of their actions to determine if they are solving or improving the problems identified.

Currently, the political process is identifying the problems or needs for COLAs. This is a legitimate way to identify problems for consideration by policy makers. However, the problems addressed in this manner may or may not be those that an overall policy framework would address.

Planning Can Identify Problems. Problems addressed through public policy can also be identified internally, through planning efforts. However, we found no evidence of an internal effort, either within the Legislature or the WRS, to assess the need for COLAs. During the past 15 years, interim committees of the Legislature have not been studying the need for COLAs, yet the Legislature has been willing to respond to pressure for them.

We reviewed the May 1996 WRS strategic plan and found it to focus on internal process operations rather than upon the external considerations that prompt COLA requests. Moreover, board meeting minutes we reviewed did not document discussions of what the board hoped to accomplish with COLAs.

Although a September 1996 version of WRS' strategic plan incorporates some planning features, it appears this is a new initiative. The board has traditionally placed its priority on developing a solid financial base for the system, and has acquiesced to the Legislature's COLA initiatives. The board has the long-standing expertise and access to information needed to create informed policy related to COLAs. To this point, however, there is no indication the board has offered or the Legislature has requested that this expertise be applied to the development of a comprehensive COLA policy.

Recommendation: The Legislature and the board should work together to develop a policy for granting COLAs.

In developing COLA policy, the Legislature and board should broaden their focus beyond affordability to other factors that have an impact in this area. These include environmental factors such as the political, economic, and demographic conditions in the state and among system members. Further, they should assess the groups of people who will likely be the focus of this policy, and the competing interests among those groups.

This recommendation acknowledges the board's stated goal to increase the cap on the automatic COLA to three percent and to oppose flat-rate ad hoc adjustments. However, our review did not indicate that this goal evolved from a process that either involved the Legislature or took into account all of the factors noted above. In any event, the board itself seems to have recognized that it has not been able to generate much legislative interest in this goal. Participating in policy development would give the board an opportunity to present its rationale and perhaps generate the desired interest.

By offering a defined benefit plan, the state has implied a commitment to public employees' income security in retirement. A COLA policy can make that commitment more explicit by specifying what it intends to accomplish through cost-of-living increases. Therefore, in developing a COLA policy, the Legislature and the board need to determine what constitutes an adequate retirement income and how much of a retiree's initial purchasing power should be maintained.

Since there are no national standards for benefit programs, Wyoming would not be the first to grapple with this issue. Nonetheless, some states have developed policy guidelines that better enable them to make decisions with respect to COLAs. Both taxpayers and WRS members would be well served if Wyoming policy makers defined a policy and designed a strategy to meet the particular needs of this state.

Finding 2: Enhanced Management Information Could Aid the Decision-Making Process

Opportunities exist to enhance the kinds of information available to legislators and board members who are responsible for considering cost-of-living adjustments under the WRS. This is important because professional literature suggests that decisions concerning pension benefits and funding are only as good as the information on which they are based.

We found that some potentially useful information was not being tracked or disclosed to either the Legislature or the board. Many of the system's reports include extensive amounts of information, but omit such significant factors as the impact of inflation on benefits or the allocation of investment earnings that exceed assumed rates of return. Although some current and former board members we interviewed were generally satisfied with the materials provided, others expressed concerns that program literature was too technical and did not always address their information needs. It appears that much of the available material is primarily oriented toward managing the system's investments and is not very helpful in developing an understanding of how inflation has affected pension purchasing power.

Prior Adjustments Have Not Been Analyzed

WRS managers told us they have never done a comprehensive analysis of cost-of-living adjustments to determine the extent to which COLAs have maintained pension purchasing power. In order to analyze this type of information, we randomly selected the pension records of 125 retirees. Our sample included both recent and long-term retirees. Using information from the system's computerized database and some original records, we determined original benefit amounts and dates of retirement for this group. We then compared their current adjusted benefits to the amount they would have received if their benefits had been fully indexed for inflation.

For the sample we reviewed, we concluded that the cumulative effects of automatic and ad hoc increases provided to retirees have substantially helped to maintain pension purchasing power. Our analysis showed the median purchasing power for retirees in our sample was 96 percent (based on 1996 dollars). In other words, the median purchasing power had declined by just four percent when looking at this group as a whole.

Although our review showed the median purchasing power had declined by only a small amount overall, we also found that adjustments had benefited some individuals more than others. As shown in Figure 5, the pension purchasing power for individuals within the group varied from a low of 77 percent to a high of 738 percent.

Many of the individuals with the highest percentage gains were retirees who had earned their pension under the old money purchase plan and therefore had the lowest original benefits. For example, the individual with the 738 percent increase had received an original monthly benefit of $7.88, based on 81 months of service. Over time this individual's pension had been increased to $217.55 per month. The results of our analysis were consistent with many comments we heard that the Legislature has been responsive in granting ad hoc increases for older retirees who received the poorest benefits.

Additionally, we found there were more individuals who had lost purchasing power than who had gained. Of the 125 records analyzed, 77 individuals (62 percent) lost purchasing power, while 48 individuals (38 percent) gained purchasing power.

Because WRS managers have not developed a mechanism to track pension purchasing power over time, it has been impossible for the board or the Legislature to systematically target cost-of-living adjustments to groups most in need of benefit increases. This lack of purchasing power analysis has made it difficult for the board and Legislature to respond to advocates of further pension adjustments. Further, it has contributed to a system that relies heavily on the legislative process for granting cost-of-living adjustments.

This analysis of pension purchasing power for 125 retirees provides some indications of issues in need of attention. In addition, we believe it underscores a need for more rigorous analysis of purchasing power by WRS managers.

Better Information on the Affordability of Proposed COLAs is Needed

WRS managers, board members, and legislators have traditionally based their decisions about proposed cost-of-living adjustments on whether or not increases are "actuarially sound." According to this principle, a pension system should be funded in a reasonable way; the principle requires sufficient funding, not necessarily full funding. While the board's approach is consistent with good management practices that require advance funding for all promised benefits, balanced decisions depend on the availability of enough useful information that can be easily understood.

We found the retirement system's process for determining the affordability of proposed benefit increases is problematic. The current method of determining affordability does not directly acknowledge "excess" earnings from investments, which occur when earnings are above assumed rates. The amounts involved can be substantial, considering the system's assets are valued at more than $2 billion and the annual assumed rate of return is eight percent. For example, in 1995 the system's investments earned about 25 percent.

Our review showed that investment earnings which exceed assumed rates have routinely been dedicated to improving the system's basic funding position. Moreover, under current procedures, information that clearly and specifically identifies the total amount of these "excess" earnings has not been provided to either the board or Legislature.

Administrators explained that the costs of proposed benefit increases are determined based on input from the system's actuary. These estimated costs are then compared to the system's available "funding margin," which is the difference between the employer's costs and contributions for a particular year. The funding margin can be positive or negative, although it has been ten years since the system experienced a negative funding margin.

Administrators reported the system's funding margin for calendar year 1995 was 0.53 percent. Based on a total payroll of $835 million, we estimate that over $4.4 million would have been available for benefit enhancements, according to the system's established process for determining the affordability of benefit increase proposals.

Advocates of benefit increases have questioned whether this method fairly considers investment earnings which exceed assumptions. The actuary confirmed that $125 million in excess earnings had been dedicated to better funding for the system between 1986 and 1994. Actuary reports also show that an additional nearly $120 million had been added to the system's assets in this way during calendar year 1995, for a total of $245 million over a ten-year period.

These practices have undoubtedly contributed to an improvement of the retirement system's fiscal soundness in recent years. Actuary reports show the system's funding ratio (assets as a percent of accrued liabilities) improved from 64.9 percent to 95.3 percent between January 1984 and January 1996. Also, this process has helped to eliminate the system's historic unfunded liability more quickly than had been planned. The unfunded liability is effectively being amortized over a 13-year period from January 1996; this is significantly less than the board's current funding policy, which stipulates a 28-year amortization period from the same date.

While these are positive outcomes, they have occurred without the deliberate actions of either the board or the Legislature. We believe the lack of explicit information on excess earnings has precluded the board and the Legislature from the opportunity to consider policy alternatives, such as COLAs, other benefit enhancements, or reductions in contributions.

Others Have Recognized a Need For Management Information

The Wyoming Legislature has taken some steps to recognize the need for reliable program information related to pension benefits provided to retired public employees. W.S. 9-3-405 requires the board to file with the Legislative Service Office statements describing proposed benefit changes and the estimated cost of such modifications. Another section of state statutes, W.S. 9-3-419, suggests an interest on the part of the Legislature in keeping track of cost-of-living as it relates to pension benefits. Public employee retirement systems in some other states, such as Idaho, California, and Colorado, regularly produce management reports that include the information needed to support board decision making on benefit adjustments.

Recommendation: WRS managers should develop more understandable and useful information related to COLAs.

Examples of information that could be helpful include an analysis of pension purchasing power and total amount of excess investment earnings. We recommend that WRS prepare management reports that clearly identify all investment earnings that accrue to the system over assumed rates, or perhaps even establish a separate fund within the system to receive these revenues. First steps in the development of improved management reports would likely involve polling other retirement systems to identify useful information and formats, and discussing these options with board members and legislators themselves. Managers should also consider the information needs of the public and members of the system.

Finding 3: A More Structured Process Could Facilitate Legislative Consideration of COLAs

The Legislature has not established a standard process or a formal mechanism for systematically handling benefit enhancement proposals and other retirement-related issues. This factor has contributed to the lack of coordination between the Legislature and board on COLAs. Moreover, without its own formal process for considering retirement issues in an integrated manner, the Legislature is less able to work together with the board to anticipate problems and develop effective solutions. Coordination between these parties is vital because of the importance of the retirement system to public employees, the size of the assets at stake, and the complexity of the policy issues.

A subtle tension exists between the Legislature and the board. The Legislature faces pressure from constituents to grant benefit improvements, while the board attempts to be fiscally prudent with assets. The roles the two parties have adopted each have merit, but have not been effectively communicated and reconciled.

The Legislature and Board Need to Effectively Coordinate Roles

The principal contact between the board and Legislature occurs during the legislative session, with WRS staff representing the board's point of view. Traditionally, the Joint Appropriations Committee considers retirement matters which have budget impact, while other types of retirement legislation are heard in different committees. This fragmented process does not facilitate seamless decision making at the legislative level, or between the Legislature and the board.

While understanding the pressure legislators face, several board members believe the Legislature's historic authorization of ad hoc increases has not been in the best interests of the system. One stated, "Legislators are responding to a very real constituency, and responding with their hearts. But at the same time, this practice tends to shift the responsibility which was given to the board...."

The Legislature Could Benefit From Board Expertise

The Legislature and board do not have a tradition of jointly anticipating and planning for COLAs. Historically, benefit advocates have approached individual legislators to sponsor COLA legislation. Thus, legislators have tended to react to individual issues as they arise, without the full benefit of the board's expertise.

Our interviews and review of board minutes suggest there is not a strong history of advocates approaching the board to request COLAs, prior to seeking a specific legislative solution. As a result, the board does not have an opportunity to reach an accord with advocacy groups to anticipate and incorporate changes in the system. As a former board member described it, "Requests for improvement come about without the understanding of the costs and desirability of funding them. Then they are decided by political pressure rather than an analysis of what's wise or fair."

Board members have expressed frustration that advocates circumvent the board to seek changes through the legislative process. The board is forced to react to legislative efforts, and cannot plan to finance ad hoc increases in advance. The nature of the current process effectively excludes the board from legislative deliberation. Another board member said the Legislature "doesn't use the board as a resource."

A High-Level Legislative Forum Does Not Exist

The current process for handling retirement-related issues does not include an appropriately high-level forum for debate and policy development between the Legislature and board. According to our review of pension literature, at least 21 states have some type of reviewing entity responsible for pension issues, in addition to their retirement boards. These states have recognized that a broad forum is necessary for ongoing policy development.

The structure and responsibilities of these bodies vary. Some states have designated permanent or interim legislative committees, while others bring together legislators, government workers, and the public as a task force. These entities provide oversight, reduce political pressures, and offer expertise in understanding complex pension issues.

The National Conference of State Legislatures (NCSL) and the American Legislative Exchange Council (ALEC) advocate an elevated legislative role in pension policy and administration. According to NCSL, the creation of a knowledgeable legislative body is vital to effectively supervise pension plans. A designated legislative committee is better positioned to recommend reforms that reflect sound principles of pension policy, rather than give isolated responses to pressures from constituents. Professional literature notes that states with legislative commissions have successfully created a core group of legislators with special expertise in the complicated issues of pension administration.

Recommendation: The Legislature should create a high-level forum to facilitate coordination with the board.

Requests for COLAs are not likely to diminish in the future. Of the respondents to our retiree survey, 66 percent believe the COLAs they have received are not adequate. A heightened focus within the Legislature's own structure could help produce more systematic interaction and better coordination of policy. In creating such a forum, the Legislature could choose from several alternatives:

• The Legislature could establish a permanent select committee dedicated specifically to retirement issues, or it might determine that a subcommittee of the Joint Appropriations Committee is an appropriate location for such a responsibility.
• Similar to the State Building Commission, the board could invite two legislators to participate in their deliberations. More formally, statutes could be amended to include legislators on the current board. Either arrangement could increase legislative representation in board decisions, and facilitate the flow of information back to the Legislature.
• The Legislature may wish to expand the scope of the state employee compensation commission, as created in W.S. 9-2-1019, to include ongoing review of pension issues.

Retirement system staff and the board should also explore ways to encourage advocates to approach the board initially with concerns, before approaching individual legislators for increases.

Finding 4: The Board Needs to Identify Achievable Ways To Fund COLAs

Judging from the history of legislative actions and from the board's goal of raising the cap on the automatic COLA to three percent, we concluded the Legislature and board are in agreement that some post-retirement adjustments are needed. If so, methods for funding them need to be identified.

The board has not presented the Legislature with a realistic, achievable proposal for addressing the negative effects of inflation on retirees' purchasing power. Specifically, in the five years since the one-percent COLA went into effect, they have not proposed a method that identifies the funding sources necessary to accomplish raising the cap to three percent. Based on our review of other states' policies and practices, we found the board can be more effective in achieving its goal, without compromising its fiduciary responsibilities.

Increases Must Be Actuarially Affordable

W.S. 9-3-419(b) requires the board to be responsible for ensuring that any increases to the cap on the automatic COLA are actuarially affordable. Members told us that they hope to pay for COLAs by improving funding margins through increased investment returns, but that actuarial reports show the additional two percent is not yet affordable. WRS estimates the annual recurring cost of each additional one percent to be $16.8 million.

The board has not committed to a time frame for achieving its goal. It has no target date for raising the cap on the automatic COLA by two percent, and the actuary told us under the current system it would be "decades" before that could occur. The director said it might be 2002 before even a one-percent raise in the cap would be affordable. Thus, the likelihood of the system being able to afford the board's goal without a change in funding or policy appears small.

Until recently, the board had been waiting until an additional full-percent increase was deemed actuarially sound. At an April 1996 meeting, the board voted to approve a more incremental approach. The board now supports raising the cap on the COLA by one-half percent as soon as actuarially feasible, perhaps in 1997. If affordable, however, it appears the increase would be so small as to be barely noticeable in a retirement check. For example, for a retiree receiving $500 per month, the increase would be $2.50 per month. Even when compounded over future years, retirees may see this as a negligible amount and may continue to press the Legislature for ad hoc increases.

Consequences of Ad Hocs Raise Other Questions

Since legislators have not been given a realistic plan or target date for accomplishing the board's goal, they have faced a difficult choice. They could either take no action, or take actions that would be inconsistent with the board's goal. The Legislature chose the latter approach and approved ad hoc increases in 1989, 1991, and 1994. During that time, the board did not propose, and the Legislature did not enact, an increase in the automatic COLA.

The Legislature has not been made aware of certain consequences of those actions. For example, we believe some purchasing power inequity may exist within the system. If there is inequity in the system now, it will only be perpetuated and increased as each year's automatic one-percent COLA is factored in. Also, the longer retirees go without receiving an additional COLA, the more powerful becomes an intergenerational issue. Implementing the board's goal would mean those now receiving benefits must do without additional cost-of-living increases so those in the future can receive a COLA with a higher cap. These issues should be explored and discussed.

Other States Provide Models

In order to maintain some degree of retiree purchasing power, several other states have put systems in place to target resources to the funding of COLAs. For example, Minnesota has established a post-retirement investment fund as a vehicle for granting adjustments. Wisconsin has a fixed-annuity reserve which provides post-retirement increases when investment experience warrants.

In addition, several states use a planned approach to granting post-retirement increases. Our research indicates that different philosophies can guide the planning and funding for COLAs designed to preserve purchasing power.

One approach includes a sequenced process, with layers of available funds used for different purposes. One layer could be used to raise the benefits of certain individuals to a minimum level and if funds remain, they can be spread proportionately over all remaining retirees. Minnesota's version of a sequenced plan provides first for an increase that matches the CPI up to a specified percent; second, an ad hoc increase is added, based on the fund's investment gains for that year.

A different approach is to identify each retiree's deficiency in purchasing power and reduce it proportionately. For example, as described in a recent paper produced for the Government Finance Officers Association:

If the purchasing power of those retired 20 years ago has declined by 50 percent, while that of those retired five years ago is down 10 percent, improving each group by half of its deficiency would lower the older deficiencies to 25 percent and the newer to 5 percent. This type of arrangement might be particularly attractive with limited available funds where it is felt necessary to provide some form of increase to every retiree, but still direct the largest adjustment to those who need it most.

Finally, several state systems grant a COLA dependent upon investment earnings when they are in excess of the actuarially assumed investment return rate. A variation is to grant such a COLA as can be provided by a designated portion of the investment gains of the system. In these cases, the increase will not occur unless the investment yield of the system is adequate to pay for the COLA. Idaho provides a partially automatic, CPI-based annual COLA with a one-percent floor, capped at six percent. COLAs above one percent are given at the discretion of the board, based on investment earnings and the system's funding status.

Many States Do Not Operate Under The Constraints Of A Legal List

One method of generating more investment income in order to fund COLAs is to give the investing authority (in Wyoming's case, the board) fairly wide latitude in selecting appropriate investments. This enables investment managers to respond quickly to changing market conditions and thereby can enhance investment earnings. However, Wyoming's retirement statutes limit investments in preferred and common stocks to 35 percent of the book value of the system's investments. As a result, nearly two-thirds of the assets must be invested in fixed-return vehicles, which often produce lower earnings than equities.

Wyoming is one of numerous states with such constraints, which are intended to prevent fiduciaries from making risky investments. However, all states including Wyoming also adhere to a prudence rule based on the Employee Retirement Income Security Act of 1974 (ERISA). Prudence standards call upon system trustees to discharge their duties with care, skill, prudence, and diligence.

The states are about evenly divided between those which operate under a legal list and those which do not. Twenty-four of the 50 state plans operate under a prudency standard only, and do not have a legal list. For example, Wisconsin has no legal list for its retirement system because of concerns that a list could be subject to political pressures and could inhibit prompt action in changing market conditions.

With fewer or no legal list constraints, it may be possible to achieve greater return on investment. According to WRS, the system earned an additional $262 million in the ten years after it began investing more heavily in equities. Our analysis substantiates that claim. We used the system actuary's statement of the portfolio's level annual rate of return over the 11 years from 1985 to 1995, which was 10.7 percent. Assuming that an additional ten percent of the total fund value could have been shifted to equities, we found the additional earnings for each of the 11 years would have averaged $10.2 million.

The Board Can Take Steps

If the board supports an increase in the cap on the automatic COLA, it needs to identify a means of funding. We suggest three ways, in addition to seeking legislation to liberalize the legal list, in which the board could take action to establish funding for such increases:

• The board could specifically designate that all or a portion of investment earnings which exceed actuarial assumptions be used to fund COLAs. Current practices allow excess earnings to be automatically dedicated to reducing the system's unfunded liability and to reducing the amortization period more quickly than board policy requires.

• The board could change one or more actuarial assumptions, if an experience study indicates this would be a sound policy decision. In the late 1980s, the board increased the assumed rate of return on investments from 7.5 percent to 8.0 percent. Based on this change, actuarial approval was obtained for an automatic COLA capped at one percent, and it was enacted by the Legislature. WRS is preparing to undergo an experience study in 1997.

• The board could seek an increase in the statutory contribution rates. Wyoming's current contribution rate of 11.25 percent is low, compared to several neighboring states. For example, Idaho's contribution rate is 16.98 percent, Utah's is 13 percent, and Montana's is 13.4 percent. Like Wyoming, all three states provide Social Security in addition to retirement system benefits.

Recommendation: The board should adopt specific methods to pay for any desired COLAs.

The system appears to have sufficient capacity to sustain some additional COLAs, should the Legislature wish to grant them. However, the board has not demonstrated that the current contribution rate is capable of supporting a larger automatic COLA than the current one percent. If the board wishes to do more than endorse the concept of additional automatic COLAs, a planned and orderly process for accomplishing that goal is necessary.

The statutory design of the system, with legal list investment constraints and a static contribution rate, somewhat limits the board's options to identify funding. Nevertheless, the board does have authority to make substantive policy decisions regarding designation of excess earnings, setting of actuarial assumptions, and determining its own legislative agenda. Changes in any of these areas might in part or in full provide the funding to support additional COLAs.

If the board requests an analysis of retirees' purchasing power, it may wish to use the results in re-examining its goal of raising the cap on the automatic COLA. With a view to improving purchasing power equity, the board could identify existing and potential inequities and recommend that the Legislature remedy those by the strategic use of ad hocs. This approach, designed to bring retirees to a level of parity, could be a preliminary step, after which the Legislature might choose to further increase the size of an automatic COLA.

Since the WRS began as a teachers' retirement program in 1943, it has evolved and adapted to meet changing circumstances. The result has been a stronger system. The board and its professional staff have worked within legislative guidelines to significantly improve the system's financial position. As it matures, the system is in a solid position to deliver promised benefits without burdening future taxpayers. At the same time, the Legislature has granted COLAs, both ad hoc and automatic, that appear to have done much to maintain the purchasing power of system retirees.

Nevertheless, these positive outcomes seem to have been more coincidental than the results of a comprehensive strategy. The focus of our report is less on what decision makers have accomplished in this arena than it is on the lack of an established policy and process to guide those decisions. Much of the system's COLA legislation was a patchwork effort, apparently enacted in response to political pressures. What little COLA policy can be inferred has been created by default, through scattered pieces of legislation.

Given its stable funding position, WRS is at a crossroads: The Legislature and board need to work together to determine the purpose of COLAs. Now is an appropriate time to clarify policy and develop strategies to ensure a future as good as the past. Without such guidance, there is no assurance that the system will fare as well in the future.

Ultimately, the Legislature bears responsibility to citizens and beneficiaries for actions taken with respect to COLAs. We believe the Legislature can be more effective if it collaborates with the board to identify desired outcomes for COLAs. In this report, we recommend that policy makers establish a direction and agree on the purpose for COLAs.

Public interest in pension plans that are equitable both to beneficiaries and taxpayers requires a broad forum for policy development. Therefore, we recommend that the Legislature consider establishing a formal process to coordinate its actions with the board. The Legislature stands to benefit from a structured relationship, since the board and staff have technical understanding of retirement system issues and the best access to system information.

Also, we recommend that staff develop more comprehensive information to be used in decision making. Finally, we recommend that the board identify timely and achievable ways for the system to fund benefit improvements.

There are no national standards for benefit program features. Each state must grapple with its own problems and develop its own solutions. Unlike states with underfunded pension plans, Wyoming undertakes COLA policy and strategy development from a stable position. The purpose of these recommendations is to offer suggestions as to how the state can better coordinate existing expertise, capitalize on its strengths, and create an equitable and feasible approach to COLAs.

Reports competed since 1995 are available free on the Internet at http://legisweb.state.wy.us. Due to technical limitations, the format of reports on the web site have been altered somewhat to be compatible with the Wyoming Legislative Service Office's web site. The agency response, certain graphics, attachments, and appendices to these reports are unavailable on an on-line basis. Complete printed copies of program evaluation reports are available for purchase from the Wyoming Legislative Service Office, 213 State Capitol, Cheyenne, Wyoming, 82002, (307)777-7881.