The ongoing crisis in teacher pension funding has led states to consider various reforms in plan design to replace the traditional benefit formulas, based on years of service and final average salary (FAS). One such design is a cash balance (CB) plan, long deployed in the private sector, and increasingly considered, but rarely yet adopted, for teachers. Such plans are structured with individual 401(k)-type retirement accounts, but with guaranteed returns. In this paper I examine how the nation's first CB plan for teachers, in Kansas, has played out for system costs, and the level and distribution of individual benefits, compared with the FAS plan it replaced. My key findings are: (1) employer-funded benefits were modestly reduced, despite the surface appearance of more generous employer contribution matches; (2) more importantly, the cost of the pension guarantee, which is off-the-books under standard actuarial accounting, was reduced quite substantially. In addition, benefits are more equitably distributed between short-term teachers and career teachers than under the back-loaded structure of benefits characteristic of FAS plans. The key to the plan's cost reduction is that the guaranteed return approximates a low-risk market return, considerably lower than the assumed return on risky assets.

The value of pension benefits varies widely, by a teacher's age of entry and exit. This variation is masked by the uniform rate of annual contributions, as a percent of pay, to fund benefits for all. For the first time, we unmask that variation by calculating annual costs at the individual level. In California, we find that the value of a teacher's benefits ranges from about 4 to 22 percent of pay, and exhibits some idiosyncratic patterns, as is endemic to traditional pension plans. The variation in individual cost rates generates an extensive but hidden array of cross-subsidies, as winners receive benefits worth more than the uniform contribution rate, and losers receive less. Almost two thirds of all entering teachers, past and present, are losers in California. By contrast, a prominently invoked study finds that nearly all active teachers are winners there. That result is shown to be highly skewed by excluding the losses of prior entrants who left early, thereby violating the funding fact that the gains and losses of winners and losers must offset each other. Our main policy conclusion is that cash balance plans can rationalize or eliminate the current system of cross-subsidies and provide the transparency lacking in traditional plans.

The authors analyze the Arkansas teacher pension plan and empirically gauge the behavioral response to incentives embedded in that plan and to possible reforms. The pattern of pension wealth accrual creates sharp incentives to work until eligible for early or normal retirement, often in one's early fifties, and to separate shortly thereafter. We estimate the effect of pension wealth accrual on teacher separation decisions using a new longitudinal data set of Arkansas teachers and find a significant impact. We then simulate the response to eliminating early retirement and raising the service requirement for normal retirement. We also simulate a shift to a constant accrual retirement plan. The response to both reforms is complex, as some would leave earlier and others stay longer. A constant accrual plan smoothes the pattern of retirement behavior as individuals tailor decisions to their own preferences instead of those built into the pension formula.

While it is generally understood that defined benefit pension systems concentrate benefits on career teachers and impose costs on mobile teachers, there has been very little analysis of the magnitude of these effects. The authors develop a measure of implicit redistribution of pension wealth among teachers at varying ages of separation. Compared with a neutral system, we find that often about half of an entering cohort's net pension wealth is redistributed to teachers who separate in their fifties from those who separate earlier, and we also identify some variation across six state systems. This implies large costs for interstate mobility. We estimate that teachers who split a thirty-year career between two pension plans often lose over half their net pension wealth compared with teachers who complete a career in a single system. Plan options that permit purchases of service years mitigate few or none of these losses. It is difficult to explain these patterns of costs and benefits on efficiency grounds. More likely explanations include the relative influence of senior versus junior educators in interest group politics and a coordination problem between states.

This article examines the pattern of incentives for work versus retirement in six state teacher pension systems. We do this by examining the annual accrual of pension wealth from an additional year of work over a teacher's career. Accrual of wealth is highly nonlinear and heavily loaded at arbitrary years that would normally be considered mid-career. One typical pattern exhibits low accrual in early years, accelerating in the mid- to late fifties, followed by dramatic decline or even negative returns in years that are relatively young for retirement. Key factors in the defined benefit formulas that drive such patterns are identified along with likely consequences for employee behavior. The authors examine efficiency and equity consequences of these systems as well as options for reform.