As thoughts turn to the Electoral College, I want to mention a point that’s been less discussed about suggestions that states shift to awarding their electors either proportionately to the vote in that state or, as Maine and Nebraska currently do in part, by congressional district.
If all states awarded their electors proportionately, the outcome in the Electoral College would much more closely mirror the national popular vote. States have the power to do so constitutionally, but there are many practical reasons states are not going to do so. But one less commonly recognized problem is that this would significantly increase the possibility of an election having to be decided in the US House. If a third-party candidate received even a small percentage of the vote in a large state, that candidate would receive a proportionate number of electors (CA has 54 electors, so 1.8% of the vote would entitle a candidate to one elector). That factor makes it more likely than under the current system that no major-party candidate would have a majority of votes in the Electoral College.
This could be countered by requiring candidates to cross a certain threshold before they were eligible to receive any electoral votes — much as PR systems require parties to win, say 5% of the vote, before they get any representation. But recall that Ross Perot won close to 19% of the vote in the 1992 election. So any threshold would have to be set quite high to avoid this risk. And of course, we don’t know how candidates and voters would change their behavior under proportionate allocation. Maybe more credible third-party candidates would run, in hope of being a kingmaker in the electoral college. Or maybe voters would become aware of these risks and less willing to vote, as many did in 1992, for candidates like Perot.
This same issue also presents a problem, though perhaps less substantial, if more states were to consider allocating electors by congressional district. Those districts would still be plurality winner systems, and it’s not hard to imagine that in at least a few congressional districts, a third-party candidate might prevail. Consider the members of Congress at the extremes of either party; would third party candidates for President be the plurality winner in some of those districts? Remember George Bush won in 2000 with 271 electoral votes. So in a tight race, it would not take many third-party winners in congressional districts to run the risk of throwing the election to the House (and ranked-choice voting would not help here, if the third-party candidate was the majority preference in the district).
There are other issues with either of these alternatives as well. But I wanted to flag this point because I’ve seen much less discussion of it when ideas like these are raised.
Update: I just checked and Perot won 30% of the vote in Maine — his best state.