The release of the Eleventh Annual Public Pension Review, expected Thursday, November 21, will likely only add to the chorus of investors, policy analysts, and academics calling for change in the management of public pension systems. Loop's report, one of the most comprehensive and long-standing geared to the needs of the municipal credit analyst, will show scant improvement over the past complete year of data that is available (from June 30, 2011 to June 30, 2012), few highlights, and continued system wide deterioration in the 247 state plans covered by the Review. In a small ironic twist--the 77 plans covered by the Review, a small subset of the total number of local plans nationally-generally fare slightly better than the state plans. While state plans continue to garnish considerable concern from taxpayers, municipal professionals, and policy analysts; local plans have drawn increasing scrutiny given the lower tax base and limited financial flexibility available to fund them. It should be noted, though, that the 77 plans reviewed by Loop are plans culled from 20 large cities, and are not a statistical valid sample of all local pension plans.
Before we discuss some of the results, it's important to emphasize that in the world of public pension analysis, the timeliness of the data is a major problem. The data presented in the upcoming report is pulled directly from the Comprehensive Annual Financial Reports (CAFRs) of states and cities of the Fiscal Year ended June 30th, 2012. While the Fiscal Year June 20, 2013 has concluded, the majority of that financial data has not been released as of yet. Additionally, some of the CAFRs rely on actuarial reports used by the state or city from periods prior to the close of the June 30, 2012 year. Thus, in a small number of cases, we are using June 30, 2012 data from a CAFR that relied upon a June 30, 2011 Actuaries' Report as the basis for what is presented in their CAFR. Dealing with these significant time lags is a longstanding problem faced by municipal finance professionals. Given the fact that investment returns were buoyed by rising stock and bond prices, and some pension reforms were beginning to take hold over this period, the performance that has take place from June 30, 2012 to June 30th, 2013 will likely show at least some modest improvement. We won't know that for quite some time, however, due to these long financial disclosure periods.
A further complication of drawing strong conclusions from the data is the fact that states and cities do not, as of yet, anyway, disclose pension fund data in a way that allows for strict comparability. The myriad assumptions chosen to apply to the calculation of funded status and actuarially required contributions (ARC) are different from state-to-state and are subject to change. As shown in the "Plan Review Detail - State Pension Plans" section that starts on page 25 of our 44 page report, the investment return assumptions, inflation assumptions, salary increase assumptions show significant variation from plan to plan, as do a number of the more arcane technical choices plan administrators must make regarding the remaining amortization period of the liabilities, the amortization method, the number of years used in smoothing investment returns, the plan type, the Cost of Living Agreement (COLA), and the cost method, etc. Adjusting 247 state plans (and 77 local plans) to attempt to achieve better comparability is not only virtually impossible, inordinately time consuming and expensive, and fraught with complex technical calculation issues; but would serve only to shift the argument away from the fact that the data is not strictly comparable to a debate over how the data was adjusted. The goal of achieving an adjusted dataset that the majority of professionals accepted as strictly comparable would never happen despite the heroic effort.
Unfortunately, the recent Government Accounting Standards Board (GASB) pension accounting changes about to come on stream will do nothing to improve the data comparability issue and everything to make it worse. In an apparent attempt to please everyone, GASB has pleased almost no one, and certainly has not contributed to an overall solution to the major problems facing the presentation and disclosure of pension data. The new system suffers from at least three horrific defects: (1) The same plan can have different discount rates that are used over time as plan results change; (2) Different plan sponsors can have different discount rates; and (3) Results must be disclosed without investment smoothing, which will create greater volatility in year-to-year reported results.
The first two changes make the use of panel data impossible. Social scientists, when appropriate, want to be able to compare entities through time and against each other. Comparing data through time is called time-series analysis. Comparing data across entities at a single point in time is call cross-sectional analysis. Anyone who has been following the public pension problem in the U.S. will recognize immediately why analysts need panel data, which is able to accomplish both analytical approaches. Unfortunately, the GASB changes ensure that the analyst will be able to do neither. As a barometer of just how bad the GASB changes are, Moody's wasted no time in devising their own system for evaluating public pensions that relies on income and balance sheet data to generate performance criteria on paying for and managing pension fund liabilities rather than leaning on pension-related data as proscribed by the new GASB regulations.
Turning to some of the results that are highlighted in the Loop Eleventh Annual Public Pension Review:
- On a weighted average basis, the funded status for the 247 state pension funds declined by two percentage points. The number of states with improvements in their weighted average funded status was confined to a precious few.
- Thirty-one states continue to struggle with making their full Annual Required Contributions due to the continuing trend of reduced federal support for states and ongoing pressure on state budgets from slow revenue growth.
- Because of the time lag involved in realizing savings from pension reform measures, the overall decline in funded status from 2011 to 2012 does not provide a sense that the legislative progress in reducing pension costs is being translated into a reduction in state fiscal stress as of yet.
- Moreover, the progress previously made by Rhode Island and some other states is being unwound by controversies over the use of alternative investments, legal challenges to pension reform legislation, and resurgent union campaigns to stop pension reform progress.
- With new GAAP pension accounting requirements coming into effect, the confusion about the performance of pension plans is likely to increase. The fact that Moody's, has decided to create their own system of pension evaluation rather than rely on the new GAAP underscores the growing confusion over how best to evaluate these systems.
- The tepid recovery, the ongoing pension crisis in Illinois, Chicago, Detroit, and the severe underfunding in Puerto Rico suggest that public policy will remain guided by the severe pension problems of the few instead of the performance that most states are experiencing.
If there is a glimmer of hope, it is provided by events outside the timeframe of the data. Investment returns have been quite strong since June of 2012. Investment returns that boost asset growth faster than the growth in liabilities improve funded status and reduce required contributions. The significant pension plan benefit changes undertaken by over 40 state governments take time to begin to produce savings (there are up-front costs to some reforms, like switching to a defined contribution system). These savings should begin to build measurably and show results when data for June 30, 2013, is released in upcoming months. Finally, with bond portfolios representing a significant component of pension assets, the rise in interest rates we have seen already, and the prospect for further increases, will boost cash flow to plans whose funding exceeds their current level of payouts. Furthermore, the academics that calculate the pension burden using the risk-free rate to discount the future stream of liabilities will be disappointed to see a fall in the value of the those discounted liabilities, reducing the perception of the magnitude of the overall problem.
In conclusion, this year's results provide little new hope, but next year's results may.