Candidate positioning or limited choice? Voter–candidate fit in the 2024 US election

When voters say neither major-party candidate represents them, the usual diagnosis is that the candidates stand in the wrong places. There is a second possibility: voters combine their views in so many ways that no pair of candidates could fit them much better. I test both with the 2024 ANES, using seven issues on which respondents placed themselves and both candidates. Setting Harris and Trump aside, I search for the two platforms that would sit closest to the same voters. Harris and Trump, as voters placed them, come close: the best pair found removes only 8.8% of the mismatch voters faced, and moving Trump alone gets most of that. More than nine-tenths of the mismatch remains, and it lies in the number of options. A centrist third candidate, the usual remedy, removes 18.4%. Letting all three options move does better: the best three-option menu removes a further 15.7% of what the best pair leaves, about 23% in all. Its middle option is cross-cutting: relatively conservative on defense and liberal on the environment. Under plurality, electing that middle option would fit voters better than electing either outer one. Better-placed candidates would help little; the larger problem is having only two.