One more post about the electoral college

Joe Biden has been in office for almost a month now, but there’s still enough interest in the strange and archaic process that put him there to justify another post about the electoral college. This one is prompted by a recent comment I saw on Facebook, in which someone asked what would have happened if the electoral college vote in each state was proportional rather than winner-take-all.

In other words, if you keep the electoral college, including its weighting in favor of small states (by counting the total number of their senators plus representatives), but have each state allocate its electoral votes in proportion to the votes the candidates received there, what happens? Would Donald Trump have done any better?

It’s a very interesting exercise (you can try it yourself with the official returns here). The short answer is that the result is much closer: in the actual election, Biden won by 306 to 232 in the electoral college; in the proportional version, those numbers are 276 to 262. Trump doesn’t win, but he picks up thirty electoral votes.

But although Biden’s win looks much narrower, it’s also much more secure. A one per cent swing to Trump from that result would have gained him an extra five electoral votes: one each in Arizona, California, Indiana, Kansas and West Virginia. That’s still a Biden victory by four votes. But in the actual election, a swing of only half that much would have won Trump the election, swinging Arizona, Georgia and Wisconsin – even though he would still have been behind in the popular vote by some five and a half million votes.

The fact that the proportional version is closer has nothing in particular to do with Trump vs Biden; it’s a general characteristic of proportional systems. Consider, for example, the last landslide result, when Ronald Reagan buried Walter Mondale in 1984. Reagan carried 49 states and won with 525 to 13 in the electoral college. In a proportional version, that would have been 320 to 218: still decisive, but much less one-sided.

Proportional voting in the college would make it much less likely that a candidate could lose the popular vote but win the election. But it wouldn’t rule it out; the bias towards small states would still be there, as would the random unfairness associated with any district system. If you take last year’s result and assume a swing of two per cent towards Trump, he would still have been some 800,000 votes behind. But in a proportional electoral college he would have had just enough to win, 271 to 267.

Sometimes elections are very close. In 2000, Al Gore led the popular vote by about half a million votes over George Bush Jr. If you calculate a proportional electoral college vote on just a two-party basis it comes out as a tie, 269-all.* But in fact Green candidate Ralph Nader had sufficient votes to win three seats in the college (two in California and one in New York); he would have held the balance of power, and could have thrown his support to Gore or forced a vote in the House of Representatives.

Similarly in 2016: third-party candidates, with at least four seats, would have held the balance of power between Hillary Clinton and Trump. And if such a system were actually adopted, that sort of thing would almost certainly happen more often; with the chance of winning some seats, at least in the bigger states, candidates from outside the two major parties would be in a much better position to attract support.

Voting systems change voter behavior. They also, of course, change candidate behavior: proportional voting would end the situation where most states could be ignored because they are safe for one or other candidate. Instead of all the attention focusing on a few swing states, it would be distributed across the country. Marginal seats could be found in safe states, in ways that would be hard to predict: who would have guessed beforehand, for example, that West Virginia would be the closest state in allocating seats between Biden and Trump?

There would still be a few states that were foregone conclusions – no-one, for example, would bother much with North Dakota, which would reliably split two to one for the Republicans. But citizens of safe states, particularly large ones like California and New York, would suddenly find that their votes mattered.

So while it would still have some quirks, a proportional system would remedy most of the problems with the existing electoral college. And unlike abolition of the college, it could be accomplished without constitutional amendment. There is nothing to stop any state legislating now to distribute its electoral college votes proportionally, as Maine and Nebraska already do in a very imperfect fashion.

The problem, however, is that it will usually not be in a state’s interests to do that. As long as other states are using winner-take-all, moving to proportionality just reduces a state’s influence; for the change to make sense, they need to all do it together. And constitutionally it is a state decision: Congress has no power to tell them how to choose their electors.

So a shift to a proportional electoral college, while simple in theory, would in practice require a degree of bipartisan agreement that today seems almost unthinkable. And if by some chance such a consensus could be achieved, it would probably make more sense to marshal it in support of an amendment to abolish the college altogether.

* Technical note: that’s using the official returns for Florida, which gave Bush a wafer-thin lead, and therefore splits Florida’s (then) 25 electoral votes in his favor, 13-12. Unofficial counting subsequently showed that the plurality in Florida actually voted for Gore, and on that basis he would have won the national tally 270-268.