Public pension plans: myths and realities for state budgets.

INTRODUCTION

Public employee pension issues are high on the policy agenda in many states. There are, of course, traditional funding problems. But there are newer issues as well, such as attempts to move from defined benefit to defined contribution plans and possible responses to new Government Accounting Standards Board rules that require the disclosure of the costs of promised post-retirement health care benefits. There is also an important reciprocal relationship between the funding of public pensions and state finances. State pension systems are placing increasing demands on state revenues, while pension funds, because of there pre-funded nature, sometimes serve as a source of ready cash when legislatures need to balance their budgets.

In this paper we explore the interaction of state pension systems with state finances. (1) We begin by describing the nature of defined benefit plans, the funded status of public plans in the U.S., and the significance of underfunding. The pro-cyclicality of both state revenues and investment returns means that states will need to make large contributions to their plans when they are least able to do so. We illustrate this link using data from the last 17 years and find that changes in pension assets are among the most important sources of funding for state governments. We also find that, in spite of widespread underfunding, states have the capacity to manage their pension obligations effectively over time, but they face incentive problems that impede these efforts.

In several states, there are discussions of moving subsets of public employees from a defined benefit (DB) pension plan to a defined contribution (DC) plan, as in the private sector, as a way to cut expenses. However, the debates frequently cloud, rather than enlighten, the issues. In the third section, we address issues to consider in moving from defined benefit to defined contribution plans in the public sector. We find that many of the differences featured in the debate between these two plan types are imaginary--or at least not intrinsic features of the two approaches. The two substantive differences between DB and DC plans for public employees and for state budgets are (1) in the assignment of investment risk; and (2) in moral hazard problems in funding. We conclude that there are real benefits to maintaining pensions in the DB form if participants are more risk averse than state governments and if moral hazard problems can be overcome.

In the fourth section, we explore the connection between plan demographics, plan features, state finances, and pension funding with a panel of 85 state public pension plans. We find some evidence of actuarial assumption manipulation to reduce funding pressure, and that plan demographics and state tax revenues have a significant influence on funding ratios. The concluding section reviews the findings in the paper and suggests ways to provide stronger incentives for states to use their existing capacity to deal with pension problems more effectively.

STATE REVENUES AND PENSION PLAN FUNDING

State (and local) public pensions are predominately of the defined benefit form. Under defined benefit arrangements, employee and employer contributions plus the investment returns on these contributions are intended to cover the cost of benefit payments. Benefits are typically related to years of service and some measure of final salary level. In a fully funded system, the accumulated assets at any point should be sufficient to meet the benefits earned by participants up to that time if the actuarial assumptions of the plans are realized. A system is said to be fully funded if the existing assets are sufficient to cover the discounted present value of the benefits earned to date by participants, i.e., the system's liabilities. The estimation of liabilities in a DB system depends on a number of key assumptions, including the future growth of wages and the returns on investments. Small changes in these assumptions can have important impacts on the estimates of a system's fiscal health. We discuss these assumptions in more detail in the third and fourth sections. Plan sponsors, not the employees, bear the risk of fluctuating investment returns and are responsible for any shortfalls. In practice, many state pension systems are not fully funded because of insufficient contributions by governments in the past, the ad hoc expansion of benefits and, less often, the failure to meet actuarial assumptions. In fact, underfunded pension systems constitute a hybrid of prefunded and pay-as-you-go arrangements. In most states, underfunded liabilities are either explicitly (through constitutional or statutory guarantees) or implicitly the obligation of state governments. (2)

State finances are directly affected by the annual required pension payments made by governments for their employees. It is less well understood that state finances are potentially impacted to a much larger degree indirectly by changes in the asset values of the investment portfolios of their pension systems. Figure 1, which shows the magnitude of annual pension funds from various sources, demonstrates both the size and volatility of investment returns since 1993. Over the period 1993 to 2004, investment returns were in total 3.3 times larger than employer contributions and 5.8 times larger than employee contributions. However, investment returns have been subject to large year-to-year variations.

[FIGURE 1 OMITTED]

Stock market movements over this period included an unprecedented run--up of asset values in the 1990s, a precipitous decline at the end of the stock market bubble in early 2000, and a strong recovery starting in 2003. This volatility in the equity markets coincided with a similar cycle in state government income tax receipts caused by capital gains realizations as well as bonuses and stock options. The late 1990s was a period of substantial revenue growth for state governments (as well as for the federal government). From 1995 to 2000, state tax revenues grew at an annual rate of 6.2 percent and income tax receipts grew at a 7.5 percent rate. This was followed in fiscal years 2002 to 2004 by the most significant decline in state revenues since at least the Great Depression. Both the revenue increases and declines were more pronounced than predicted because of the difficulty in forecasting capital gains. More recently, revenues have begun to expand again at a strong pace (but more modestly than in the 1990s), bolstered again with capital gains.

Giertz (2003) notes that changes in pension assets have a huge, but often little-noticed, impact on state finances. He demonstrates that the relatively high levels of pension funding in the late 1990s were an artifact of the strong assets markets. These high investment returns allowed some states to reduce or eliminate their annual contributions to pension funds--just as some companies reduced their contributions to their fully funded pension funds. A few states whose plans were fully funded on an actuarial basis considered closing or did close their DB plans to new entrants, replacing them with a defined contribution plan (Papke, 2004, 2007).

Changes in asset values are of significant size compared to state revenues. Figures 2 and 3 show the magnitude of pension asset changes as a percent of total state tax revenues and individual income tax revenues. The figures show not only the size, but also the volatility of pension asset changes. Note that in four of the 12 years, the increase in pension assets exceeded state income tax collections. But there were also two years (2001 and 2002) when the change was negative and one year (2003) when it was close to zero.

[FIGURES 2-3 OMITTED]

Quantitatively, the changes in pension assets are among the most important sources of funding for state governments, but this impact is largely ignored by most policymakers and observers. Over the entire period, investments returns were 23 percent of total state tax revenues and 69 percent of individual income tax revenues. It should be noted, however, that while there is considerable annual volatility, there is more stability over the longer term. The existence of longer-term stability is a key issue that relates to both the health and structure of state pension systems. This will be discussed below.