The Governmental Accounting Standards Board (GASB) still appears to be on track for a final decision by the middle of 2012 on its proposed changes to governmental pension reporting and accounting rules contained in the Exposure Drafts issued earlier this year, but there are some signs that changes in response to the results of the field testing process, and the concerns of cost-sharing plans, may yet be possible. In the meantime, two recent studies provide contrasting views of the likely results of the proposed changes as they now stand. The first, a working paper by Robert Novy-Marx, concludes that under the GASB rules as proposed, a governmental pension plan can improve its funding status “literally by burning money.” The second, a brief produced by the Center for Retirement Research at Boston College (CRR) finds that employers and plan administrators should be prepared for funded ratios reported in their financial statements to decline sharply under the new GASB rules, but that “[p]olicymakers should not let new numbers throw them off course.“
Current Status of GASB Proposals
In response to its Exposure Drafts containing proposed changes to the accounting and financial reporting for pensions by plan sponsors and to the financial reporting for pension plans by the plans themselves, GASB received 645 comment letters, including the letter prepared by NCTR, NASRA and NCPERS, which was signed by over 130 executive directors, administrators and trustees of 109 State and local government retirement systems.
In addition, GASB held three public hearings at which 33 organizations and individuals testified. Three user forums were also held, with 19 participants, including representatives of citizen groups, bond rating agencies, academics, financial analysts and research organizations.
Finally, field tests were held. These were designed to obtain information about one-time and ongoing preparation time and costs associated with the proposals in the Exposure Drafts; to identify difficult-to-understand provisions of the Exposure Drafts; and to receive feedback on the methodology employed by field test participants on certain proposals in the Exposure Drafts. Field test participants were also asked to provide pro forma note disclosures and required supplementary information and details of the calculation methodologies used to determine the discount rate—specifically, their projections of benefit payments and plan net position.
In a memo to the GASB board members, staff summarized the results. According to this memo, GASB received results from 18 participants, including 3 single-employer Plans; 3 single employers that also report plans; 3 multiple-employer agent plans; 2 agent employers; and 7 multiple-employer cost-sharing plans.
As noted, participants in the field test were asked to provide estimates of time and cost related to both initial implementation efforts and ongoing efforts that would result from the Exposure Draft proposals. (For purposes of the field test, it was requested that the estimates of time include only internal staff time and that cost estimates include out-of-pocket expenditures (such as consulting fees), software acquisition, and system changes but not include costs associated with staff time.) The results varied dramatically. Initial implementation time ranged from 2 hours for one agent employer to 56.000 hours for one cost-sharing plan.
Similarly, implementation costs varied tremendously, from a low of around $3,000 for one cost-sharing plan to as high as $18 million for another cost-sharing plan. This was then compared to current costs under existing GASB rules, which were much, much lower, underscoring the impact of the proposed GASB changes.
The staff memo also details many of the specific concerns with various aspects of the Exposure Draft that participants identified, as well as the methodologies used in the calculation of (1) the weighted-average expected remaining service life of active employees and (2) the proportionate share of collective totals of employers in cost-sharing plans.
The memo concludes with issues that arose from the field tests and which “will be considered in the redeliberations of the Exposure Drafts,” including:
- Concerns about increased internal costs and external actuarial and audit costs, particularly in multiple-employer plans in which employers have different fiscal year-ends. GASB staff notes that “Specific concerns were expressed regarding the valuation of assets without market prices and the determination of other changes in plan net position that currently are evaluated only annually”
- Applicability of the proposals to hybrid plans and plans in which defined contribution pensions may be converted to annuities
- Request for additional guidance related to proprietary funds and entities that use regulatory accounting
- Request for clarification regarding the projection of benefit payments in circumstances in which employees provide services to multiple employers that participate in a plan or in which employee contributions are greater than service cost
- Request for additional guidance related to plan reporting of investment types, pensions of the plan’s own employees, and presentation of a statement of plan net position
- Concerns about proposed plan note disclosures, including issues related to the separation of investment expenses from returns in the calculation of return on plan investments, difficulty in obtaining or calculating rates of return on a money-weighted basis, and a general concern about the volume of disclosures.
All of this suggests that these “redeliberations” may well result in changes to the Exposure Draft, although it should also be noted that , based on the projected GASB work plan, the final voting on the Exposure Draft is still scheduled for the May-June, 2012, time frame. In other words, the “fast-track” still appears to be in place.
One area of the proposed changes in the GASB rules that has been particularly worrisome has been those related to the manner in which cost-sharing plans will be required to allocate their net pension liability, pension expense, and deferred outflows of resources and deferred inflows of resources to their participating employers. The impact of this aspect of the proposed GASB changes on employers who participate in cost sharing plans could be devastating.
For example, in a recent article on the specifics of the cost-sharing proposal as it applies to employers, David Powell with the Groom Law Group stresses that “If these proposals are finalized without change, participating employers will need to substantially revamp their accounting systems at considerable expense.” Compliance is expected to be the responsibility of the employer, he explains, “and the retirement system's ability to provide assistance in many cases may be limited.” Powell warns that this “may leave the burden of compliance on the contributing government employers” and not the plans – a point that many plan sponsors have yet to fully appreciate.
NCTR continues to believe that cost-sharing plans constitute a form of insurance and that the costs should not be allocated to the individual employers, and has made this point consistently in comments filed with GASB. Nevertheless, for now, GASB appears to be set on its allocation approach.
Accordingly, a number of cost-sharing plans have formed an ad hoc group, the Employer Cost-Sharing Coalition, which has filed separate comments with GASB focused solely on this issue. These comments, prepared by the Groom Law Group, argue that:
- There is a lack of a clear relationship between the arbitrary proportionate share and actual liability
- Varying and uncertain rules in different jurisdictions means that actual liability of a cost-sharing employer for plan underfunding, however measured, will often be difficult or impossible to predict with any certainty; thus, it is not a faithful representation of the potential liability.
- Instead, plan underfunding and other information could be disclosed in notes with a description of any rules in the jurisdiction for determining employer liability for underfunding.
- Comparability to the new FASB multiemployer rule also justifies this approach. (In 2010, FASB had initially proposed a rule similar to the one GASB has proposed, to be applicable to private sector multiemployer plans, but changed its mind because of cost concerns; because the actual payment in the future might be limited by legal constraints; and because the estimate of a liability did not represent the amount an employer would actually have to pay.)
This new approach, which has reportedly been received with much interest by GASB staff, as well as the comments from the field testing conducted by cost-sharing plans, may be having an impact. For example, GASB has now indicated that, at its March, 2012, meeting, it will consider the creation of a working group to assist the Board with regard to cost-sharing matters.
Hope springs eternal.
Novy-Marx Working Paper
Robert Novy-Marx is an Assistant Professor of Finance with the Simon Graduate School of Business at the University of Rochester, as well as a member of the National Bureau of Economic Research, which published his new paper. It is entitled “Logical Implications of GASB's Methodology for Valuing Pension Liabilities,” and in it, Novy-Marx argues that GASB does not really provide a valuation method. He notes that a valuation method should recognize that “more is more,” in the sense that adding a dollar to any given set of assets and liabilities increases the set’s value. Furthermore, he says that a valuation method should also assign a unique value to a given set of assets and liabilities.
But the GASB methodology for accounting for net pension liabilities “does not satisfy either of these conditions,” he asserts. GASB instead gives different “valuations” for the exact same assets and liabilities when they are partitioned differently among plans, he argues, and moreover, since the marginal valuation of assets can be negative under GASB, then in such cases, GASB would allow a plan to improve its GASB funding status literally by burning money.
For example, he claims that GASB, while recognizing that cash and bonds are valuable assets, nevertheless “penalizes a plan for holding these, by forcing it to recognize a larger liability.” Thus, he says, destroying a dollar (i.e., reducing its cash or bond holdings by a dollar while holding all other asset holdings fixed) reduces a plan’s assets by exactly a dollar, but can reduce its GASB liability by more than a dollar. “In these cases plans can reduce their GASB recognized underfundings by destroying assets,” he concludes.
Finally, Novy-Marx insists that GASB’s methodology for accounting for liabilities has an equivalent alternative formulation. “The explicitly prescribed procedure, which entails discounting a plan’s liabilities at the expected return on its assets, is completely equivalent,” he argues, “to one in which a plan’s liabilities are discounted at rates that reflect the liabilities own risks, but the plan’s stock holdings are valued at more than twice their market values.”
The CRR brief is entitled “How Would GASB Proposals Affect State and Local Pension Reporting?” It takes a look at the GASB proposals and focuses on those elements dealing with valuation of assets and liabilities, namely (1) plan assets would no longer be smoothed but rather valued at market; (2) liabilities would be discounted by a blended rate that reflects the expected return for the portion of liabilities that are projected to be covered by plan assets and the return on high-grade municipal bonds for the portion that are to be covered by other resources; and (3) the entry age normal/level percentage of payroll would be the sole allocation method used for reporting purposes. CRR examines these effects by first determining funded ratios based on current GASB standards and then funded ratios calculated using the market value of assets. Next, CRR combines market assets with liabilities discounted by the blended rate to demonstrate the full impact of GASB’s proposed changes.
Looking at the effect of the change from an actuarially smoothed l value of assets to a market value, CRR finds that the aggregate funded ratio using market assets was only 67 percent in 2010 compared to 77 percent using actuarial assets. “[P]olicymakers should be prepared for a sharp decline in funding if GASB introduces this change,” the CRR brief concludes.
Next, looking at the new blended rate for determining liabilities, CRR found that the aggregate funded ratio of state and local plans (based on 2010 numbers) would have decreased from 77 percent to 53 percent under the GASB plan.
Thus, CRR finds that the GASB proposals “will sharply reduce the reported funded levels of public sector plans.” However, CRR also stresses that it would be “unfortunate if the press and politicians characterized these new numbers as evidence of a worsening of the crisis when, in fact, states and localities have already taken numerous steps to put their plans on a more secure footing.” The Brief concludes that “[r]eforms need to be done carefully and thoughtfully, remembering that pensions are an important part of the total compensation of public sector workers.”
The CRR brief also makes a number of other observations that are very similar to those made by NCTR and its members in connection with the GASB exposure drafts. For example, CRR says that there will be implementation problems, the main one being with GASB’s proposed blended rate. It will require a complicated calculation based on a number of assumptions, including assumptions not only about plan returns but also about future contributions from the government and from employees. “These contributions may or may not come to pass,” CRR observes, and “[o]ne can imagine extended disputes about the validity of the underlying assumptions.”
CRR also criticizes the new GASB discounting proposal because the data it will produce will fail to provide meaningful measures of government obligations and be inconsistent across states and localities and over time.
Finally, the CRR brief is worried about the impact on the annual required contribution, or ARC. First, the move from actuarial to market value of assets and the new liability measure increase the unfunded liability and thereby the required amortization payment, CRR notes. Then, a blended discount rate will raise the normal cost. Therefore, CRR cautions that “reported” ARCs are likely to increase substantially but employers may likely continue to use the traditional actuarial smoothing techniques to calculate their ARCs for funding purposes. As NCTR has warned, there could be an unfortunate and confusing disconnect between the reported number and the number being used for funding.
Next, CRR is also worried about what they refer to as the “disciplinary role of the ARC” and the likelihood that it will be undermined. Specifically, they point out that in states with statutory contribution rates, they will no longer technically be required to calculate an actuarial ARC. “This change not only represents a loss in analysts’ ability to assess how close plan contributions are to those required to keep the system on track,” CRR notes, “but also creates an escape valve that states could use as ARCs rise beyond reach: introduce a statutory rate and dispense with ARC calculations.”
In an effort to address some of these concerns with the loss of the ARC, NCTR and other national public sector organizations, in conjunction with the public plan actuarial community, are holding preliminary discussions to determine what could be a possible alternative and how its uniform use could be encouraged.
One final note regarding the CRR brief: appendix B of the brief contains a “run-out” date for 126 public pension plans. It is well to note that CRR has reassured Keith Brainard, Research Director of the National Association of State Retirement Administrators (NASRA), that these dates are not actual projections of when the plan could become insolvent. According to CRR, these dates “are not reflective of an ongoing plan, and are only intended for use in implementing GASB’s particular liability concept.” As Keith puts it, “in other words, these dates are based on an accounting basis, not a funding basis.”