Chad Aldeman at Bellwether Education Partners tweeted that my recent report on teacher pensions was “frighteningly bad.” Here’s what’s really frightening: a zombie lie about teacher pensions that won’t die.

Attacks on teacher pensions may not rank with global warming or mass shootings on the list of things keeping us awake at night, but they deserve way more attention than they get. Teacher pensions are the single most important tool for recruiting and retaining good teachers, and good teachers are the key to our future in a knowledge economy. Yet Aldeman is trying to mislead people into supporting Kentucky Governor Matt Bevin and others around the country who want to switch teachers to 401(k)-style plans.

Aldeman claims that “most teachers get a bad deal from teacher pensions.” Though my research, and earlier research from UC Berkeley, showed that the vast majority of teachers are well-served by their pensions, this recycled claim appears impervious to counter-evidence. As long as a billionaire with an agenda keeps funding the research, we’ll continue see elaborate variations on the same theme, all of which rely on the same statistical sleight-of-hand.

The average person understands “most teachers” to mean “most teachers teaching today” or at any given point in time. This is a commonsense interpretation, and the one used in my research. Aldeman’s methodology instead counts new teachers as they enter the system, giving them equal weight whether they teach for just one year or for a whole career. As I showed in my paper, even if slightly over half of new teachers leave before becoming eligible for employer-provided benefits, these short-term teachers, some of whom go on to earn pension benefits in different systems, represent only a tiny fraction of the teaching workforce.

Aldeman accuses me of callously ignoring the “lived experiences” of individual teachers to focus on a “snapshot” that ignores “anyone who was once a teacher and is no longer (a teacher).” This is nonsense. Aldeman is no more focused on individual teachers’ lived experiences than I am. He and I rely on the same experience studies based on the same pension participants—including teachers who leave. The only difference is how we weight participants. He gives equal weight to all new teachers who enter the system in a given period, and I give equal weight to all teachers active at any given point in time.

To understand the difference, it might help to take a step back from teachers and pensions and imagine a school in a rural area with a military base. Eighty percent of students in the school come from farm families, and twenty percent are children of military personnel. Advocates of eliminating a school 4-H program and replacing it with a ROTC program claim that “most” students come from military families and have no interest in 4-H. Their evidence? Counting all students who ever enrolled in the school and noting that more than half came from military families. Since military families don’t tend to stay long, however, these students represent half of new enrollees but only a fifth of the student body at a given point in time.

Substitute “teachers” for “students,” “pensions” for “4-H,” and “401(k)-style account plans” for “ROTC” and you see how Aldeman and other pension critics twist language and use misleading statistics to give the impression that “most teachers” would be better off in an account-style plan even when four out of five active teachers are career teachers on track to earn full benefits. Some of the remaining teachers—a small fraction of whom leave before becoming eligible for benefits, plus a somewhat larger group who receive full benefits that are tied to lower mid-career salaries—might be better off with 401(k)-style plans. Whether they actually would be depends on a number of assumptions, such as whether teachers become eligible for benefits more quickly in the account-style plan.

You may wonder why it has to be all or nothing. Why not have pensions and account-style plans? There’s nothing wrong with adding a supplemental retirement savings plan—that’s what 401(k)s were always meant to be. But Governor Bevin and other would-be reformers are trying to eliminate pensions altogether based on the lie that most teachers don’t fare well under the current system.

Governor Bevin seems to think that he can cut compensation to underpaid teachers without any effect on recruitment and retention. Public school teachers in Kentucky, three quarters of whom have a master’s degree or more education, earn $47,000 on average, much less than similarly-educated private-sector workers in the state.[1] Nationwide, teachers earn less than private-sector workers with similar experience and education, even accounting for summers off.

Traditional pensions aren’t expensive—dollar for benefit dollar, they’re more cost-effective than 401(k)s, where retirement outcomes are eroded by high fees and low investment returns. Pensions do get expensive when elected officials neglect contributions for years, but these legacy costs don’t go away just because you switch plan type for new employees.

Governor Bevin wants to gut teacher benefits at a time when Kentucky schools face a struggle to hire qualified teachers. Enrollment in teacher preparation programs in Kentucky has dropped by more than half in just two years as teachers are asked to do more and more for less and less. Switching to 401(k)s would just exacerbate, not reverse, this trend.

[1] Source: EPI analysis of inflation-adjusted 2012-2016 U.S. Census American Community Survey data for Kentucky teachers working 27+ weeks per year and 20+ hours  per week and earning $10,000+ per year via IPUMS-USA