More generally, the argument in the peer-effects literature is that we shouldn't let smart kids escape the public school system because their presence gives dumb kids a positive externality. I detest this argument. Children are not pawns to be moved about to satisfy the desires of some grand master. A decent school system treats children as ends in themselves.Characterizing the benefit of smart kids to dumb kids as a “positive externality” raises an interesting point. Let’s suppose it’s true that the presence of smart kids really does improve the educational performance of the dumb kids. The primary lesson of the Coase Theorem is that externalities can often be internalized through market transactions. Thus, to use one famous example, the fact that honey bees provide positive externalities to apple orchards (by pollinating apple blossoms) does not mean there will necessarily be an inefficiently low level of beekeeping. Why not? Because the owners of apple orchards can pay the beekeepers to release bees in their orchards. And indeed, this is exactly what happens, according to a 1973 article (“The Fable of the Bees”) by Steven Cheung.
What is the analogous arrangement in schooling? Scholarships, of course. Lower-ranked colleges regularly make sweet offers to good students to lure them away from higher-ranked schools, because such students are likely to raise the school’s prestige and make larger contributions as alumni. The non-scholarship students benefit from both effects. If it’s really the case that smart students create positive “peer effects” externalities, there’s every reason to believe that private schools in a voucher system would find ways to achieve the optimal mix. And they would do so without treating the best students as merely means to an end, because they would have to reward such students for their beneficial choices.