HD52 - Review of the Virginia Retirement System

  • Published: 1994
  • Author: Joint Legislative Audit and Review Commission
  • Enabling Authority: House Joint Resolution 392 (Regular Session, 1993)

Executive Summary:
The Virginia Retirement System (VRS) administers a statewide public employee retirement system which provides defined benefit pension plan coverage for State employees, teachers and non-professional employees of public school boards, and employees of participating political subdivisions. In addition to the State system, VRS administers separate retirement systems for State police officers and judges, a group life insurance program, a deferred compensation program, and a health insurance credit program for eligible State retirees.

Currently 222 State agencies, 146 local school divisions, and 353 political subdivisions participate in the retirement system. In addition, 132 local school divisions include their non-professional employees in the system. At the close of fiscal year 1993, VRS had 259,086 active members, and 86,369 retired members, inactive vested members, and beneficiaries. Total pension fund assets were valued at $15.9 billion. Retirement benefits paid in FY 1993 totaled $667.9 million.

House Joint Resolution 392 of the 1993 Session of the General Assembly directed the Joint Legislative Audit and Review Commission (JLARC) to study the structure, investment policy, and actuarial soundness of the Virginia Retirement System. The impetus for this study grew out of concerns raised about the independence of the VRS and about the soundness of some investment decisions made by the Board of Trustees (Board). To complete the review, JLARC staff examined the structure and governance of the retirement system. In addition, professional investment and actuarial consultants assessed the soundness of VRS investments and funding.

Structure and Governance of the Virginia Retirement System. VRS has experienced tremendous asset and membership growth in recent years. The phenomenon of rapid growth has transformed the management and operation of VRS into one of considerable complexity. Consequently, there are now greater demands on the governing structure of the retirement system.

Recently VRS' system of governance has been called into question. Concerns have been raised about the appointment of trustees to the VRS Board, as well as the qualifications for Board membership and the independence of the Board as a governing body. Similar questions have been raised concerning the Board's two advisory committees. Furthermore, several recent Board actions have diverted attention from substantive issues associated with the soundness of the system, and have created a negative public perception of VRS, especially among State and local employees.

All Appointments to VRS Board are made by the Governor. Virginia is one of only eight states, all of which have retirement systems with smaller total assets than VRS, in which the Governor has the sole power to appoint retirement system trustees (see figure on page ii). This has contributed to a perception among VRS members that the Board is not entirely independent of the executive branch in its decision-making. Some of the appointments made by recent governors, such as cabinet members, have contributed to this perception. In addition, complete gubernatorial appointment authority does not properly reflect the General Assembly's constitutional responsibility for the retirement system. The appointment of trustees would better reflect the responsibility of the General Assembly and improve the independence of the Board if some trustees were appointed by the Legislature.

Recommendation (1). The General Assembly may wish to amend Section 51.1-109 of the Code of Virginia to require the General Assembly to appoint some members of the Virginia Retirement System Board of Trustees.

Qualifications for Trustees Are Inadequate. In recent years, oversight of VAS investments has become the Board's most prevalent, and time-consuming responsibility. However, the required qualifications for serving on the VRS Board have not kept pace with this growing responsibility. Current statutory requirements for membership on the VRS Board tend to focus on representation of specific types of VRS members rather than on professional qualifications of the trustees.

According to JLARC's investment consultant, Bear Steams Fiduciary Services, the issues involved in the area of public pension fund investment require informed judgment and significant expertise at the Board level. Bear Steams concluded that a majority of VRS trustees should have experience in the investment of large employee benefit funds.

There are other problematic aspects of the current qualification requirements for Board membership. One trustee, while designated as the political subdivision employee representative, is actually an elected local constitutional officer. Two other trustees are appointed to their positions of employment in State government by the Governor, and serve at his pleasure. This situation promotes a perception of undue gubernatorial influence on the Board, and raises questions about the Board's independence.

Recommendation (2). The General Assembly may wish to require that a majority of Virginia Retirement System trustees have experience in the direct investment of large funds. Representation for teachers, State classified employees, and local employees should be continued.

Recommendation (3). The General Assembly may wish to amend Section 51.1-109 of the Code of Virginia to prohibit elected officials and executive branch appointed officials from serving on the Virginia Retirement System Board of Trustees. However, the State Treasurer could be appointed as an ex-officio, non-voting member.

Perceptions of Board Actions have Eroded Confidence. Since 1990 the VRS Board has been involved in a series of events which, at least at first glance, call into question the independence of the Board and its ability to effectively govern the system. These recent events have detracted from the public image of VRS. For example, at a public hearing held by JLARC, a member of the Virginia Governmental Employees Association (VGEA) stated that, "in recent years a cloud has been placed over the retirement system by controversial actions of the VRS Board, thus the confidence of the beneficiaries of the system has been eroding."

Among the issues which have raised concerns about the retirement system are the RF&P Corporation acquisition and subsequent appointments to the RF&P board of directors, compliance with the Freedom of Information Act, the Redskins stadium proposal, and public disputes over investment policy. All of these events have left an impression of a Board which is influenced by political considerations, which is unnecessarily secretive, and which is unable to effectively govern the retirement system. While these impressions may not all be based on fact, the perceptions continue to exist among many members of the VRS.

The Role of the Board Chair Needs to Be Redefined. Over a long period of time, the position of chair has acquired a degree of perceived, but not necessarily intended, power and authority. This power and authority appears to stem from the fact that the chair is appointed by the Governor. However, such power is only implied since the chair has no statutory responsibilities. While the Board recently defined the role of the chair as part of its policies and procedures, the stated responsibilities are minimal. The Code of Virginia should set out a clear role for the chair to provide leadership for the Board and to communicate on its behalf.

Recommendation (4). The General Assembly may wish to amend the Code of Virginia to define the role and responsibilities of the chair of the Virginia Retirement System Board of Trustees.

Strong Chief Investment Officer Needed. The VRS investment staff has grown gradually with the increase in the size and sophistication of the fund. Under the current structure, the investment staff is supervised by a chief investment officer (CIO). However, since the departure of the CIO in 1990, that position has been vacant.

Most recently, the Board initiated a search for a CIO. However, according to the revised job description for the position, the CIO seems focused on management of the investment department rather than on substantive investment direction and coordination. Bear Steams reviewed the plans for the CIO and found that the position may not meet VRS' long-term needs.

Bear Steams has recommended that the VRS investment department be managed by a CIO who has overall responsibility - under the lAC and the Board of Trustees -for the organization, structure and performance of the VRS investment department and investment portfolio. Bear Steams also recommends that consideration be given to providing in the Code of Virginia for the selection, appointment (possibly via a special employment contract), and new reporting duties for the CIO at the VRS.

Recommendation (5). The General Assembly may wish to amend the Code of Virginia to establish the position of chief investment officer for the Virginia Retirement System. The duties of the chief investment officer should include coordination of asset allocation; communication with trustees, advisory committees, and the General Assembly; and staff support for the VRS Board of Trustees and its advisory committees.

Recommendation (6). The General Assembly may wish to provide for the employment of the chief investment officer by special employment contract which would set out performance and formal reporting requirements. The General Assembly may also wish to require that the appointee to the position be confirmed by the General Assembly. The employment contract should require the chief investment officer to make periodic reports to the General Assembly.

Structure and Role of the Advisory Committees Can Be Strengthened. The current advisory committee structure, consisting of the investment advisory committee (lAC) and the real estate advisory committee (REAC), appears to be a fairly well-organized and useful system. However, in order to ensure that the advisory committee structure continues to serve VRS well in the future, the committees' role and structure should be formally defined in statute. In addition, because of the importance of the investment advice provided to the Board, the necessary investment-related qualifications for advisory committee membership need to be set out in the Code of Virginia.

Recommendation (7). The General Assembly may wish to amend the Code of Virginia to require the Virginia Retirement System Board of Trustees to formally maintain an Investment Advisory Committee and a Real Estate Advisory Committee. The Code of Virginia should define the general responsibilities of the advisory committees

Recommendation (8): The General Assembly may wish to define in the statute qualifications necessary for membership on the Investment Advisory Committee and the Real Estate Advisory Committee. In addition, the General Assembly may wish to require that a majority of the members of each advisory committee meet such standards.

Independence of the Trust Fund Could Be Strengthened. The VRS pension trust fund is established exclusively for the benefit of VRS members in section 51.1-102 of the Code of Virginia. In addition, the VRS pension plan is a qualified plan under the provisions of the U.S. Internal Revenue Code. As a result, the pension trust fund is exempt from federal taxation on its contributions and investment earnings.

Despite the State statutory language and the IRS restrictions, questions have been raised periodically concerning the long-term ability of the trust funds, and the retirement system, to function solely on behalf of VRS members and retirees. For example, while federal law prohibits transfer of trust fund assets, maintenance of specific contribution levels to ensure the actuarial soundness of the trust funds is not required.

Because of concerns about the adequacy of current statutory language establishing the retirement fund as a trust, a constitutional amendment would provide a means to better define the independence of the fund. A total of 13 states have some sort of constitutional provision concerning the funding of their retirement systems.

Recommendation (9). The General Assembly may wish to consider amending Article X, Section 11 of the Virginia Constitution to include the following provisions: the VRS retirement funds are independent public trusts, the assets of which are not subject to appropriation by the General Assembly or for use as loans for other State purposes; and the financing of VRS pension benefits shall be based on sound actuarial principles, with employer contributions consistent with the recommendation of the VRS actuary.

A Proposal for Strengthened Governance. In order to ensure that VRS is properly governed as it grows into the next century, the General Assembly needs to consider a comprehensive restructuring of the retirement system. The restructuring should focus on enhancing the independence of VRS, and imposing more stringent qualifications for Board membership to better reflect the increasing complexity of retirement system investments.

Implementation of a new structure of governance for the Virginia Retirement System as outlined in the body of this report will be a complex task, involving the creation of an independent agency, the transfer of benefit programs to another agency, and the appointment of trustees who meet the new qualifications. In order to most effectively implement the new structure, the current VRS Board should be dissolved on the effective date of the new structure. Consequently, the newly configured Board and agency would constitute a complete replacement to the current system of VRS governance.

Legislative Oversight of the Retirement System. Due to the General Assembly's constitutional mandate to maintain a State retirement system in the best interest of the members, adequate legislative oversight of VRS is essential. However, the General Assembly's ability to provide effective oversight is limited, in large part, because of inadequate communications between VRS and the General Assembly. Without such information, oversight cannot be carried out effectively. To address this problem, a new process for legislative oversight is needed. Specifically, the General Assembly may want to create a permanent oversight commission for the Virginia Retirement System.

Twenty states have some type of oversight entity responsible for monitoring their retirement systems. The structure and responsibilities of these oversight bodies vary. For example, some consist entirely of legislators while others include public members. However, all of these oversight bodies provide their legislatures with independent sources of retirement system information.

Recommendation (10). The General Assembly may wish to establish a permanent Virginia Retirement System Study Commission to provide ongoing oversight and evaluation of the retirement system. The Commission should be composed of three members from the Senate of Virginia, three members from the House of Delegates, and three qualified professionals appointed by the Governor. To carry out its duties, the commission should have a permanent staff and the authority to hire consultants. Funding for the commission should be from the retirement system trust funds to ensure continuity and independence.

Recommendation (11). To ensure an effective system of oversight, the General Assembly may wish to establish the following responsibilities for the VRS Study Commission: receive quarterly and annual reports from the Virginia Retirement System on actuarial soundness and investment performance; review and report as necessary on all proposed legislation affecting VRS' structure, investments, or funding prior to the consideration by the standing committees of the General Assembly; prepare and maintain background and other information for use by members of the General Assembly; make an annual report to the General Assembly and the Governor on the status of the retirement system; and conduct special or continuing studies as directed by the General Assembly.

Investment Policies and Performance of the Virginia Retirement System. The investment performance of a public pension plan is of great importance to both plan participants and taxpayers because of the major role investment income plays in overall financing. As a result, it is critical that retirement systems develop and implement fundamentally sound frameworks to govern investment decision-making. The importance of this in Virginia is amplified by the rapidly growing size and sophistication of the State's public pension fund.

The consultants hired by JLARC to evaluate the State's retirement system indicate that the investment program and portfolio structure are fundamentally sound and reasonable in almost all major respects from both a procedural and substantive standpoint. There is no cause for concern in either the investment decision-making process or in the results of that process. However, the consultants recommended several improvements to the investment program.

Statutory Investment Requirements Need Revision. The type of investment restrictions imposed by the Code of Virginia are commonly referred to as "legal lists." These types of restrictions are fairly common, but a review of other states' statutes indicates that at least 32 states impose fewer investment restrictions than Virginia. By enacting this legislation, the General Assembly wanted to articulate and impose standards which would require the VRS Board to act with requisite care and expertise and to prudently construct and oversee a diversified investment portfolio. In its current form, however, the statute fails to achieve these goals and contains many investment restrictions which Bear Steams found to be ambiguous, inapplicable, or superfluous.

Recommendation(12). The General Assembly may wish to consider amending the Code of Virginia by adopting a prudent person standard without a legal list, comparable to the standard set forth in the Employee Retirement Income Security Act.

An Integrated Investment Policy Statement is Needed. The VRS Board is responsible for determining what objectives the fund should seek to attain in order to generate sufficient cash to pay required retirement benefits. However, the system has not adopted an overall "Statement of Investment Policy and Objectives" for the entire fund. Generally speaking, such an investment policy would reduce to writing the basic objectives and the overall framework within which all investment strategies should operate.

Recommendation (13). The VRS Board of Trustees should adopt a written investment policy statement drawing from the Five Year Plan, the Policies and Procedures Manual, and other appropriate sources.

Recommendation (14). Once this policy is adopted, the VRS Board of Trustees should re-evaluate the investment policy statement at least annually and either reaffirm or amend it as appropriate. Periodically re-evaluating the investment policy statement has the added benefit of compelling the Board, lAC and staff to continually reassess the VRS' investment objectives and the basis for those objectives.

VRS' Asset Allocation Policy Needs a More Thorough Review. Asset allocation is the process of diversifying an investment portfolio among asset classes, (stocks, bonds, cash, real estate, etc.). This is done in order to seek to achieve a particular investment objective, such as consistently earning a specified total return (i.e., income and appreciation). A portfolio's asset allocation is important because it has the single greatest impact on its overall long-term investment performance, far greater than the specific securities held in the portfolio.

Because of this, a portfolio's asset allocation policy should be reviewed and adjusted on a periodic basis as appropriate. However, as an agenda item at VRS' annual retreat this issue does not appear to get the attention it deserves simply because there are already so many other items covered at the meeting.

Recommendation (15). The VRS Board of Trustees and the lAC should review the asset allocation policy as a formal agenda item for detailed discussion at some point each year in a setting other than the annual retreat.

Process for Selecting and Terminating Managers Can Be Improved. Investment managers provide money management services for a portion of the fund's assets, for a fee, on a fully discretionary or non-discretionary basis. As of August 31, 1993, the total VRS portfolio was managed by 75 external investment managers. The processes VRS uses for selecting external investment managers generally appear thorough and based on appropriate criteria. However, some reluctance by the Board and lAC to make the difficult choices among several qualified candidates for investment management slots was observed. The consequence of this is a tendency by VRS to hire more than the required number of firms.

Recommendation (16). VRS should develop procedures to reduce duplication in the hiring and continued retention of managers, enhance the selection criteria for its money managers by adding liability insurance, increase the staff responsible for its domestic equity program, and improve its policies for determining the time period over which a manager must meet required investment objectives.

VRS Employs Too Many Investment Managers. By comparison to plans reporting in the recent PENDAT survey by the Public Pension Coordinating Council/Government Finance Officers Association, the VRS employs a high number of investment managers. Compared to a more select group of funds that Bear Stearns surveyed, the VRS also seems to have a large number of managers.

VRS Brokerage Practices Are Reasonable. VRS appears to obtain reasonable value for the soft dollars it expends and appears to pay reasonable commissions for the quality of securities execution it receives. According to an outside study, VRS' average commissions have been below the median cost incurred by other pension funds and the securities prices which brokers have obtained on trades for VRS are reasonable. According to the same study, the VRS' average transaction costs (i.e., commission cost plus execution cost) have been 5.4 cents per share, which is below the median cost of 6.3 cents per share for other surveyed pension funds.

VRS Has Developed A Properly Diversified and Efficient Portfolio. The principle of diversification is essential to the VRS portfolio. Although some newer asset classes (such as managed futures, venture capital, and international investments) taken in isolation are often considered riskier in some respects than conventional stocks and bonds, VRS concluded that these asset classes have certain attributes which, when combined with the stock and bond components, may actually lower the volatility of the total VRS portfolio and raise the expected ratio of return to risk.

The efficiency of the VRS portfolio was tested with computer simulations of various combinations of the ten subclasses of assets used by the VRS. This comparison indicates that the VRS asset allocation is fairly efficient and that the expected returns meet the VRS objectives over the long term. It was also found that VRS' asset allocation appears reasonably structured to produce satisfactory returns at a relatively low level of volatility or risk. However, as compared to the portfolio structure five years ago, the newer asset classes have probably also contributed to returns lower than those of some other public funds on an absolute basis.

Performance Meets Internal Objectives but is Less than Typical Pension Fund. Over the five-, three-, and one-year periods ending June 30, 1993, the VRS has met its own long-term (10-15 years) internal objectives of earning more than the actuarially-assumed rate of return and exceeding the rate of inflation by at least four percent per year. The VRS has not, however, met its short-term internal objective of earning at least as much as other large pension funds. VRS' total returns over the past one, three, and five years are slightly below the median returns for a broad sampling of other public funds in the widely-used Trust Universe Comparison Service. In addition, VRS returns fell generally at the low-to-mid range of returns for a smaller group of public funds selected for comparison by Bear Steams.

Performance is Comparable to Customized Benchmarks. Since VRS has a diversified, complex portfolio, Bear Steams devised several customized indices to help evaluate VRS' performance. Bear Steams calculated the risk-adjusted returns for VRS against the three customized indices over five years. The risk-adjusted returns for the VRS over the five-year period ending June 30, 1993, were approximately equivalent to the risk-adjusted returns for the three customized indices over that same period.

VRS' Managed Futures Program is a Good Diversification Tool. The structure of the VRS managed futures program is novel, and -- subject to a few important exceptions -- is reasonably well-designed to detect and control risk. Rather than contracting directly with commodity trading advisers ("CTAs") who manage futures portfolios, VRS has hired five registered investment advisors (RlAs) who, in tum, select and monitor a wide variety of CTAs. The criteria for selecting RIAs appear reasonable. However, the process does not include consideration of what, if any, errors and omissions liability insurance each RIA candidate carries. In addition, one RIA had not imposed written guidelines on its CTAs at the time of this review.

Recommendation (17). VRS staff Should review the nature and specificity of the new guidelines recently imposed on CTAs to assure that the staff is satisfied with them.

Performance of Managed Futures Program Has Met Expectations but Fees are Excessive. The performance of the managed futures program since inception has been largely as expected and satisfactory. However, the fees of the outside consultant are problematic for two reasons. First, the structure or formula for those fees embodies a potential conflict of interest. Second, the absolute total amount of fees paid appears unduly high. To the extent a renegotiation does not achieve sufficient reductions, re-bidding the futures consulting and monitoring contract should occur.

Recommendation (18). VRS should restructure its fee arrangement with RP Consulting, to base it only in minor part, if at all, on turnover. Instead, the fee should be based on the amount of equity in the program or a flat fee.

Recommendation (19). VRS should negotiate a lower fee for its futures contract. If a lower fee cannot be established, VRS should re-bid the contract.

Actuarial Soundness of the Virginia Retirement System. VRS provides competitive pension benefits to its members. In order to continue providing these same benefits, the pension fund's assets, increased by future contributions, must be sufficient to cover the cost of all future benefits. An important objective of VRS pension funding is to provide benefit security for its active and retired members, so as to ensure that promised benefits will actually be received by VRS members.

The actuarial firm which has served as the VRS actuary since 1980 has provided the VRS Board with competent and responsive actuarial services and advice. Partly as a result of this firm's efforts, accrued retirement benefits are currently adequately funded. However, the funding status of VRS could deteriorate over the long term.

VRS Funding Status Will Decline Over the Long Term. According to 30-year actuarial projections prepared by JLARC's actuarial consultant, Alexander & Alexander, the overall funding status of VRS, as measured by the ratio of assets to liabilities, will decline in coming years. The primary cause of this decline is the pay-as-you-go approach used to cover the cost of the COLA benefit. On the other hand, the value of current accrued retirement benefits is funded to a much greater extent. JLARC's consultant projects that, mainly due to the lack of prefunding of the COLA, total VRS employer contribution rates will increase significantly over the next 30 years.

Criteria for Establishing Funding Target. VRS should achieve the funding target over a reasonable period of time, such as ten years. The period should not be so short as to cause dramatic increases in contribution rates. In addition to establishing a funding target, VRS should examine projections to determine the contribution rates required to meet and maintain the target. VRS should also specify how the funding level will be held within a certain degree of tolerance of the target.

Recommendation (20). The Virginia Retirement System Board of Trustees should establish a funding target, ensure that its actuarial assumptions and methods are appropriate to achieve the target, and then monitor progress toward the target.

Recommendation (21). The Virginia Retirement System Board of Trustees should examine the long-term trends in funding status through the use at open group projections. At the same time, the VRS Board should examine the sensitivity of these projections to the assumptions of future experience.

Recommendation (22). The General Assembly, and the Virginia Retirement System Board of Trustees, may wish to consider alternative methods of funding and providing cost of living adjustment benefits.

Recommendation (23). The Virginia Retirement System Board of Trustees should identify and consider available options for the funding and provision of pension benefits to ensure that short- and long-term costs can be held to acceptable levels.

Technical Aspects of Actuarial Cost Method Cause Contribution Rates to Increase. The entry age normal cost method is generally accepted and in common use, especially among state retirement systems. A majority of the state retirement systems use this cost method.

According to Alexander & Alexander, certain technical aspects of the valuation process could result in increased contribution rates even if all assumptions are met exactly. These increases occur even if no changes are made to actuarial methods, assumptions, or plan provisions. In particular, three elements of the cost method application should be modified.

Recommendation (24). The Virginia Retirement System Board of Trustees should modify the actuarial valuation process as follows:

• Recognize the timing lag in determining the employer contribution rate;
• Reduce the amortization period for current unfunded accrued liability by two years each biennium; and
• Amortize all additional unfunded accrued liability, from plan amendments, actuarial gains and losses, and assumption or method changes, separately over a reasonable period, such as 15years, from the inception of the additional unfunded liability.

Actuarial Assumptions Are Currently Reasonable, But Need Long-Term Revision. The current economic and demographic assumptions used by VRS are reasonable, and similar to those used by other state retirement systems. VRS' approach to establishing its assumptions is in line with accepted actuarial practice. However, unidentified sources of actuarial loss in the 1992 experience investigation point to the need to reassess all of the assumptions. In addition, the long-term implications of short-term changes in actuarial assumptions need to be determined prior to implementation.

Recommendation (25). The Virginia Retirement System Board of Trustees should analyze its economic actuarial assumptions.

Recommendation (26). The Virginia Retirement System Board of Trustees should Implement the changes in demographic assumptions recommended by Buck Consultants in its 1992 experience investigation.

See full report for Recommendations (27), (28) and (29).