This illustration can only be used with the Sarah Smith story that originally ran in the 10/3/2015 issue of National Journal magazine. Richard Mia

Two researchers at Georgetown University have an idea for paying teachers more while improving students’ education, all without increasing costs for school districts. They propose redistributing students to the top teachers’ classrooms, increasing class sizes slightly, and using the money saved to pay a bonus to teachers who take on more students. They first wrote up their idea in a paper for Georgetown’s Edunomics Lab. I recently spoke to one of the researchers, Georgetown professor Marguerite Roza, about the idea. Our exchange has been edited and condensed.

—Sarah Smith

How would the proposal work?

The paper proposes offering our best teachers the chance to take on additional students and paying them more for it. Because they took on additional students, the system overall would need fewer teachers, which would create the savings to pay them the higher salaries. There has been quite a bit of research on the effects of the most effective teachers on students, and those effects seem to trump class size. What we did is we looked at the financing of it, to see how the math could work out, knowing the effects for students would be positive.

How would the money redistribution work? Would this require getting rid of any teachers?

We suggest giving the teachers the full amount of the savings. We did not say that we’re laying off any of the weakest teachers—that brings in a whole additional set of policy barriers. What we’re saying is simply to identify your top quartile and say to them, “Over time do you want more students in your classes? If so, you’ll get a higher pay bonus.” The average in the country would be a 19 percent bonus—$10,357. Every school district sees attrition on a regular basis. That would involve not hiring more teachers if your district is growing or not replacing teachers who are leaving because you don’t need to.

We thought the most immediately understandable version would be for a district that was undergoing enrollment growth. They could go: “Before we hire more teachers, let’s put those students in the classrooms with our most capable teachers. And doing so creates an immediate way to hire one less teacher.”

In the paper, you propose giving three kids to each high-performing teacher. How did you arrive at that number?

People won’t tolerate 50 kids in a class. What we wanted to do was present a nice increment of students—say, three, a relatively modest increment—and a substantial amount of money involved in possible pay raises. We picked a number we thought was still in the realm of class sizes that people could digest. Of course, schools are welcome to add even more.

Is there a point at which adding too many kids to a class counteracts what a really good teacher can do?

I don’t know that we know. Frankly, I suspect that depends on the teacher. But parents, when asked if they’d rather have a teacher with higher class size and a teacher from the top quartile, or lower class sizes and worse quality, went for the known quality even though it meant higher class sizes. But I suspect it’s a matter of district policy that people don’t want to put 50 kids in classes.

What system would be in place to check if instruction quality went down with more students added?

Probably the best answer is whatever system is in place now. Schools are already measuring their classrooms, right? And student outcomes could let them know if something is going south.

What would it take to make this happen?

It means letting go of some of your fixed-class-size, fixed-pay schemes. You don’t even need to change your salary schedule, just make sure you have a stipend structure that allows overpayment. You can’t have a fixed-class-size cap. I know in Texas, where we modeled this, that the district can ask for a waiver from the state, so it’s possible there, although the state has a class-size policy. You have to identify who your most successful teachers are. Then they always have to have an assurance to parents that they’ll watch that data on a year to year basis.

We want to hear what you think about this article. Submit a letter to the editor or write to