GG is hard to define but also has a known-when-seen quality. I’d look for papers on policy effects and/or correspondence between policy/electeds and “median voter.” Can say more. Part of the measurement issue is that coord failure of type alluded to above seems rare.