Comment by tzs

There should be a lot of natural experiments that can test this. Just find the boundaries of municipal water systems that fluoridate and take a look at the populations near the boundary but on opposite sides.

Often those boundaries will cut through a neighborhood and so you end up with a population of demographically similar households but some are on city water and some have wells.

Many of those neighborhoods will have fluoride levels in their well water that are quite different from the levels in their city water, making them good subjects for a study.