I don't remember too much from grad school because all that advanced math that cluttered up my brain was replaced with mad shitposting skills, but one thing which always stuck with me was a certain paper I read at some point. It didn't stick with me enough that I remember the authors and title, of course, but the basic empirical result.
It was an old paper and who knows how robust and repeatable it was, but this was it:
A major determinant of the likelihood of school bond issues passing in local elections was if the voting population was white and older and the school aged population was less white. That made them less likely to pass, of course.
