Teacher Pensions Blog

For an organization that claims to care about workers, the Economic Policy Institute (EPI) is awfully dismissive* of how teacher retirement policies play out for individual teachers.

In a new report for EPI, Monique Morrissey asserts that, “teachers and schools are well served by teacher pensions,” and attacks our work looking at how many teachers benefit from today’s teacher retirement plans.  

The debate centers on one critical question: Should we care about individual workers’ lived experiences in a given profession, or should we take a snapshot of the current workforce as representative of all those who experience it? How you think about that question will inform how you think about how well most teacher pension systems work today.

Morrissey wants us to consider a snapshot, or cross section, of the active teaching workforce. Today’s teachers have a range of ages and experience levels. Some are just entering the profession, while others are closing in on retirement. It’s hard to capture that diversity and Morrissey attempts to take this cross-sectional approach and answer the question of how many teachers will reach the 30-year mark. In her estimation, 80 percent of today’s teachers will do so.

But here’s the problem with that method: a snapshot of today’s teachers omits anyone who once was a teacher and is no longer. For every teacher who’s closing in on 30 years of experience today, there were many others who started out at the same time who are no longer part of the picture today. We know this because state pension systems maintain good data on participants and what they are owed. For instance, about one-third of new, young teachers in North Dakota will make it to 30 years, compared to only about one-in-ten teachers in North Carolina.  

A snapshot, then, is irrelevant to determine what percentage of all teachers will receive adequate retirement benefits, because employees accumulate retirement savings as individuals. It’s also a misleading way to analyze the data. Under defined benefit pension plans, like the ones serving most public-school teachers, teachers receive retirement benefits according to their own salary and their own years of experience. If they serve only a short period of time, they don’t get much in the way of retirement benefits. Moreover, pension plans don’t deliver benefits in a linear fashion, and they provide disproportionately large benefits to teachers who stay for a full career.

We don’t have to speculate about this or come up with hypotheticals; every state posts their data on pension plan websites.

That’s why our work on teacher retirement looks at the chances any particular teacher has of reaching various retirement milestones. And again, our figures are not derived from speculative models, we use data from the state pension plans themselves, which use the same data in their own financial calculations.

Pension plans have to estimate how many teachers will reach various stages of their career and, in turn, qualify for pension benefits. Every pension plan publishes these assumptions in their financial reports, and they conduct regular “experience studies” to see if their assumptions are correct or if they need adjustments. (Perversely, state policymakers use this data to make pension plans even harder to qualify for in order to shore up their flagging finances, for instance in 15 states it takes ten years to vest for even a minimal teacher pension; that’s a problem for workers that EPI quickly glides over).

Since Morrissey relies heavily on a prior research paper on California (which I’ve written about before), I'll use an example of California’s most recent experience study, adopted in 2017. The graph below shows the pension plan’s estimate for voluntary (“withdrawal”) by the teacher’s year of experience. This graph is for females, but the state has separate, albeit similar, assumptions for males and for other sources of turnover, like retirement, death, or disability.

In the graph, the skinny blue line was California’s prior assumption, and the bars represent the actual experience from 2016 and 2011, respectively. Based on its observations, the state is slightly tweaking its assumptions going forward, to the orange line.

Again, every state needs these assumptions to be accurate, or else it could risk over- or under-saving for future pension payments. After running their experience studies, states publish their final assumptions in tables like the one below. The way to read the table is that, for any female member in her first year of service, California expects 15 percent to leave and 85 percent to stay. Out of the remaining 85, California assumes another 9 percent will leave in year two, another 7 percent in year 3, and so on.

What we’ve done in our work is to look at a typical teacher starting out in a state and ask, if we follow the state turnover assumptions, what are a teacher’s chances of reaching the state’s vesting requirement, when she’ll first qualify for any benefit? And, what are her chances of reaching the state’s normal retirement age?

We believe those are the right questions to ask. They’re the questions any given teacher should ask in planning for his or her own retirement, and they’re the questions the pension plans are asking themselves.

The answers to those questions show that pension plans are not working well for large majorities of people who enter the teaching profession. In the median state, about half of all new teachers won’t stay long enough to receive any pension at all, and only about one-in-six will stay for a full career.  That’s not a typo. That’s how few teachers make it. 

No one is in dispute of these estimates. They look slightly different for men or for people who begin teaching at various ages, but in reality the main driver of teacher turnover, at least early in his or her career, is years of experience, not age or gender. (For simplicity’s sake, we typically only report the survival rates for new, 25-year-old females, but a more sophisticated study by Bob Costrell and Josh McGee found similarly depressing results even when looking at teachers who enter at all different ages.)

EPI doesn’t try to challenge these numbers either—on page 5 of her report, Morrissey concedes “the critics’ calculations may be accurate,” but then she tries to change the conversation to the snapshot approach. That’s a fundamentally flawed way to look at retirement security, because it discards large numbers of former teachers and ignores the basic facts about how individual teachers accrue benefits over time.

Reasonable people can disagree about what to do about this problem. Some advocate 401(k) style approaches while others favor traditional pensions but with some tweaks. And there are hybrid options like “cash balance” plans that have aspects of both of those approaches. The reality is that there are trade-offs with every approach and it’s possible to design lousy pensions and good 401(k)s and vice versa. What we shouldn’t disagree about, however, is that this conversation is too important not to be informed by the best data and honest straightforward analysis of what it means and what those trade-offs and choices are.

 *Note: This piece is mainly about EPI’s methodology, but their report also includes a number of alarming statements about teachers. For example, at one point they admit that “vesting requirements are primarily cost-saving, not retention, devices,” essentially waving away the fact that half of all new teachers won’t stay long enough to vest and sticking a knife in literally millions of American workers. Moreover, this statement concedes the point Kelly Robson and I made in a recent article for Education Next that showed that pensions are not a retention incentive for the vast majority of teachers.