Abstract
This commentary explains how changing labor market conditions have made it more important for schools to attract and retain skilled teachers and yet more difficult to do so. This is important context for considering the contributions of the five papers on teacher pensions in this special issue. I describe and comment on the evidence that defined benefit (DB) pension plans have deleterious impacts on the ability of public schools to provide students with an excellent education. I describe the merits of alternative strategies for responding to the central problem with DB plans as currently funded. I conclude by explaining the importance of considering the design of teacher pension plans as part of a multipronged strategy to provide all students with a high-quality education.
Pensions are an important part of the compensation of career teachers. Their structure and generosity affect the ability of schools to attract and retain college graduates who will become excellent teachers. Staffing the nation’s schools with skilled teachers is considerably more difficult to do today than 50 years ago because women and college graduates of color have better labor market options (Corcoran et al., 2004). The COVID-19 pandemic has placed new burdens on teachers and reduced the attractiveness of teaching relative to alternative careers. Nearly one in four teachers reported that they were likely to leave their jobs by the end of the 2020–2021 school year, compared with one in six teachers prior to the pandemic (Steiner & Woo, 2021).
Not only have changing labor market conditions in recent decades made it more difficult for schools to attract and retain skilled teachers, but they have also made it more important to do so. Changes in the economy have made cognitive and social skills more important determinants of labor market earnings than was the case several decades ago (Deming, 2017; Murnane et al., 1995). For many decades, researchers have shown that teachers have an important impact on students’ cognitive skill development (Boyd et al., 2009; Chetty et al., 2014; Hanushek, 1972; Murnane, 1975). Recent research shows that teachers also influence in important ways students’ development of social skills (Jackson, 2018). By increasing students’ cognitive skills and social skills, teachers influence students’ potential for labor market success.
In summary, changing labor market conditions have made it more important for schools to attract and retain skilled teachers and yet more difficult to do so. This is important context for considering the contributions of the five papers on teacher pensions in this issue. Three of the five papers focus on defined benefit (DB) pension plans. The paper by Andrew Biggs provides a clear explanation of how DB plans work. In the next section, I describe and comment on the evidence presented by the authors that DB pension plans have deleterious impacts on the ability of public schools to provide students with an excellent education. I argue that one of the concerns the authors raise about DB plans is much more important than the others. In the subsequent section, I describe the merits of alternative strategies for responding to the central problem with DB plans as currently funded. I conclude by explaining the importance of considering the design of teacher pension plans as part of a multipronged strategy to provide all students with a high-quality education.
Defined Benefit Pension Plans
Career teachers (defined as teachers who enter the profession early in their working lives and remain teaching until retirement) typically view DB pension plans as valuable. Such plans provide a guaranteed monthly pension that is a set percentage of the salaries teachers earn during their final years of teaching. DB plans remove financial market risk, especially if the pension payment is adjusted for changes in the cost of living. Advocates for DB plans point out that they provide strong incentives for teachers to remain in the profession until they have achieved the number of years of service that entitles them to their maximum pension (Ghilarducci, 2011). Although Goldhaber and Holden (2023) do not evaluate the relative merits of DB and defined contribution (DC) pension plans, they do provide compelling evidence that teachers are willing to give up some current income to increase their income in retirement.
The authors of the papers on teacher pensions in this issue point out a variety of problems with DB pension plans for teachers as they are typically designed. One problem is that between 1990 and 2010, gaps arose in many states between the value of the DB pension obligations to teachers and the funds put aside to pay the pensions. As Biggs (2023) explains, contributing factors were increases in the generosity of DB plans during the 1990s and the consequences of the Great Recession, which included declines in the value of investments in state pension funds and declines in state tax revenues. The shortfalls in funding to pay DBs created serious financial problems in many states. However, most states have responded in ways that have put their DB plans on sound financial footing (Biggs, 2023). Common responses have been to increase the required contributions teachers must make toward pensions and to increase the number of years of service incoming teachers are required to teach to receive a full pension (Biggs, 2023).
Mihaly and Podgursky (2023) point out that DB plans hinder the mobility of teachers across states because they typically cannot transfer all the state’s contribution toward their pension to another state’s pension plan. Economists see this as a problem because a mobile workforce helps alleviate state-specific teacher shortages. However, I do not see this as a critical problem with DB plans because most teachers teach quite close to where they grew up (Boyd et al., 2005; Reininger, 2012). Therefore, teachers are not very mobile, even within states. And, of course, state legislators see the attraction of DB plans as an incentive for teachers to continue to teach in the state where they typically were educated.
Mihaly and Podgursky (2023) point out that in a 401(k) plan, recipients may pass on unspent funds in a retirement account to heirs, but this is typically not the case under DB plans. This option is an advantage of DC plans. However, DB plans can be structured to pass on benefits to heirs if retirees die soon after retirement. For example, in the West Virginia Teachers Retirement DB plan, a retiring teacher may choose a “10 Year Certain - 120 Payments option” that pays a reduced monthly pension throughout the retiree’s lifetime. If the retiree dies before receiving 120 monthly payments, the remainder of those 120 monthly payments is paid to the retiree’s named beneficiary or the retiree’s estate (West Virginia Consolidated Public Retirement Board, n.d.).
In my view, the most serious problem with DB teacher pension plans concerns the income redistribution from short-time teachers to career teachers. As Costrell (2023) explains clearly, “Under traditional [DB] plans, the full cost of a career teacher’s benefits far exceeds the contributions designated for them.” A large part of the monetary gap is filled by money the state puts into the pension fund but does not pay to short-term teachers, who are entitled only to their own contributions plus interest when they leave the profession. The same goes for teachers who leave after vesting but short of a career unless they are willing to wait until retirement age for a pension. The number of teachers affected by these pension provisions is large: 44% of new teachers leave the profession within 5 years of entry (Ingersoll et al., 2018).
Analysts have different views on the extent to which the income transfer from short-term to career teachers is a problem. Ghilarducci (2011) sees the transfer as a strength of DB plans. Her argument is that beginning teachers are not very effective, becoming more effective during the first 10 to 15 years in the classroom and remaining effective for a significant number of years thereafter. Consequently, by providing incentives for teachers to stay in the profession, the income transfer promotes efficiency, especially if the tenure decision removes ineffective teachers from the profession. Recent research showing that teachers do continue to increase their effectiveness for at least the first 10 years in the classroom supports Ghilarducci’s argument (Kraft et al., 2020).
However, as Costrell (2023) and Biggs (2023) point out, the money that is transferred from short-term teachers to career teachers through the DB pension plan could be used to improve entry-level and early-career salaries. The importance of doing this is suggested by the modest starting salaries in teaching. In 2015–2016, the average starting salary for public school teachers with a college degree and no teaching experience was $38,800 (Ingersoll et al., 2018). Increasing early-career salaries would increase the attractiveness of teaching to college graduates with attractive occupational alternatives.
In summary, the most serious problem with DB pensions as conventionally funded is the tension between providing attractive guaranteed pensions to career teachers and using available funds to increase the pay and improve the working conditions for beginning teachers, thereby making teaching a more attractive early-career alternative. This tension is important because the percentage of teachers who remain in the profession for only a few years is high.
Policy Alternatives
A strength of the papers on teacher pensions in this issue is their emphasis on paying attention to incentives and examining teachers’ responses to these incentives. This surely is good advice. Hosek et al. (2023) and Kong and Ni (2023) illustrate how an econometric structural model can provide insights about the ways in which teachers’ responses to incentives influence the attractiveness of policy alternatives. As these authors illustrate, these models can be useful in predicting the consequences of proposed changes in teacher pension plans.
Although the authors of several of the papers in this issue provide detailed descriptions of incentive problems associated with DB pension plans, they do not provide parallel analyses of incentives and responses associated with alternatives. This is understandable because there have been very few natural experiments providing alternative pension plans for teachers.
At the same time, the authors of two of the papers in this issue (Costrell, 2023; Mihaly and Podgursky, 2023) comment favorably on the features of DC plans. I agree that DC plans, such as 401(k)s, have some attractive features—in particular, transparency and benefits to short-term teachers that DB plans do not provide. However, I do not see DC plans by themselves as contributing to the goal of attracting and retaining effective teachers. One reason is that DC plans that are acceptable to teachers and to state governments are unlikely to provide enough money for retiring teachers to purchase an annuity that would provide 70% of their salaries from their last years of teaching, the replacement rate actuaries recommend for a comfortable retirement (Biggs, 2023, fn. 3). To provide sufficient funds, either the required minimum contribution from teachers would be too high to be acceptable to many or the required matching rate from the state would be too high to be acceptable to legislators.
A second problem is that if the required minimum contribution from teachers is low enough to be acceptable, a great number of teachers would not supplement their required contributions sufficiently to fund a comfortable retirement. Clark et al. (2016) study the responses from when Utah closed its DB plan to new teachers and replaced it with two choices, both of which provided less generous benefits than the DB plan and shifted risks from the state to teachers. One choice was a DC plan. The other was a hybrid plan that combined a modest DC plan with a less generous DB plan than the DB plan it replaced. The research team finds that the majority of teachers did not make an active choice between the two plans and consequently were included in the default hybrid plan. These teachers did not participate in the supplemental DC plan.
A third problem is that a substantial percentage of teachers lack the knowledge to make good investment decisions (Levitz, 2008). Many are also unprepared to take on the investment risks that 401(k) plans entail. Yet another problem is the so-called red truck syndrome, which refers to the tendency of people receiving a large 401(k) payout when they retire to make large durable-goods purchases, underestimating how much money they will need in retirement (Levitz, 2008).
If DB plans as conventionally funded reduce the attractiveness of teaching for talented college graduates who envision teaching for a while but not for a career, and if DC plans reduce the attractiveness of teaching for those who envision a career in the classroom, what are the alternatives? One is a two-part strategy that combines modest changes in DB plans to increase their efficiency with the allocation of additional tax revenues to salary increases for early-career teachers. One change in DB plans could include increases in the number of years of teaching required for a full pension. Another could be basing pensions on the inflation-adjusted average salary over the last 5 years of teaching rather than the last two or three. This would reduce the incentive to take on extra duties to increase compensation in the last years of teaching. States that are committed to improving the quality of education provided to all students should consider this strategy, especially combined with the complementary policies described in the next section. Unfortunately, concerns with tax levels preclude serious consideration of this strategy in many states.
An alternative strategy that many states have embraced is a hybrid pension plan (Mihaly & Podgursky, 2023). These typically combine a less generous mandatory DB plan with a mandatory DC plan to which both employee and employer make monthly contributions. The short-term teacher is entitled to keep not only their contribution to the DC plan but also the employer’s contribution. However, the teacher does not keep the employer’s contribution to the DB part of the plan. The career teacher has a less generous monthly pension at retirement, perhaps replacing 45% of their teaching salary. However, they also have significant funds in their 401(k) plan that they can use to purchase an annuity. In 37 states, teachers participate in Social Security and consequently receive benefits from this national DB plan as well as from their state’s DB plan.
Important Complementarities
The five papers on teacher pensions in this issue document the challenges in designing teacher pension systems that are attractive to talented graduates who are considering teaching for a modest number of years before moving to another occupation and to graduates interested in a career of classroom teaching. Both groups are heavily represented in the teaching force, and both do a lot of the teaching of the nation’s children.
In evaluating alternative designs for teacher pensions, it is important to keep in mind that the impact of financial incentives embedded in pension designs on the quality of education is heavily dependent on other personnel policies. For example, Papay and Kraft (2016) show that hiring teachers after the school year starts reduces student achievement. Therefore, compensation policies that increase the quality of the applicant pool will have a greater impact on the quality of new hires if recruitment and screening are done early. Jackson and Bruegmann (2009) show that the effectiveness of beginning teachers depends critically on the skills of their grade-level colleagues. Therefore, placing novice teachers in schools with strong teaching faculties increases the probability that they will become effective teachers. Kraft and Papay (2014) show that the extent to which teachers improve their performances as they gain experience depends on the quality of the professional environment in their school. Taylor and Tyler (2012) show that high-quality evaluations improve teachers’ performances. Johnson et al. (2012) and Kraft et al. (2016) show that school safety and the quality of school leadership affect teacher turnover.
In summary, the authors of the papers on teacher pensions in this issue are right in arguing that the design of pensions affects the ability of schools to attract and retain skilled teachers. They are also correct in arguing that changes in teacher pension plans are needed to use available funds to provide a high-quality education to all students. However, in my view, the requisite changes can be made while retaining the DB plans that career teachers value. Finally, it is important to keep in mind that the payoff to investments in improving financial compensation, either in salaries or in pensions, will depend on the extent to which the investments are part of a comprehensive plan to recruit, retain, and develop the skills of talented college graduates.