Electoral College math experiment

With a lot of buzz going about the Electoral College and the fact that Trump won the electoral vote despite Clinton winning the popular vote, I opted to conduct a little experiment. I used the poll numbers available from the New York Times as of December 23, 2016, when this article was published. I realize they’re not the final, official numbers, but likely close enough for this experiment, and the final numbers are unlikely to change the outcomes, though I’m willing to revisit this when those numbers are readily available.

Now I’ve advocated for the Nebraska model to become universal with regard to how the electoral votes are divvied up. Nebraska I believe divides by congressional district or population, with the popular vote winner getting the two votes representing the Senate. As such, though the State almost always goes Republican, the Democrats can usually expect to pick up a vote from that State.

So if the Nebraska model were universal, and presuming they divide the electoral votes by population with the two Senate votes going to the population winner which will about represent the congressional district model, what would be the totals? Note: some rounding errors had to be corrected manually for this result, but the outcomes were not affected by the corrections.

  • Trump: 272
  • Clinton: 258
  • Johnson: 7
  • McMullen: 1 (Utah)

Trump would still win, but only just barely. And Clinton would’ve had more votes overall due to picking up votes in Texas and Florida, but would’ve lost votes in California and New York. McMullen you’ll see would’ve picked up an electoral vote in Utah, and that would’ve been due to his close run behind Clinton in that State. Trump would’ve picked up the remaining two population votes and the two Senate votes.

Also worth noting is that Clinton and Trump would’ve had votes in every State with the exception of the 7 States plus the District of Columbia that are allocated only three votes.

Now what if the popular vote was divvied up entirely by population. Would Clinton have won the electoral vote since she also won the popular vote? No.

  • Trump: 263
  • Clinton: 266
  • Johnson: 8
  • McMullen: 1 (Utah)

Clinton would’ve won the plurality, but no candidate would’ve won a clear majority. This vote result would’ve gone to the House to resolve, and at one vote per State, it likely would’ve gone to Trump.

The electoral college exists, in part, to lessen the capability of one State to control the outcome of the election for President. In both scenarios above, and in the actual outcome, that purpose is well served. Clinton’s popular vote win is fueled largely by her win in California, where she won by a larger vote margin than she did in the overall popular vote. And she won California by a vote count that surpasses the populations of about half the States in the United States.

One must also remember that the United States is not and never has been a democracy. We are a federated republic of independent, sovereign States. And the electoral vote system preserves that. The electoral vote system would stop serving that purpose if the 12 largest States all banded together to select the President, regardless of who the other States selected.

Now what about close races? Let’s look at 2000, pulling the official tallies from the Federal Election Commission. In that outcome, going by the Nebraska model, same presumption, this would’ve been the Electoral College result:

  • Bush: 275
  • Gore: 258
  • Nader: 5

The result is a little more interesting if you divide the electoral vote proportional to the popular vote:

  • Bush: 263
  • Gore: 269
  • Nader: 6

That vote would’ve gone to the House, and given the breakdown of the 107th Congress, it could’ve gone either way.

And if you really want to see how much it neutralizes the power of the largest States, let’s look at the 1972 election. In that election, Richard Nixon won 520 of the available 538 electoral votes. What would’ve been the result had the Nebraska model been in play at that time (actual total in parentheses)?

  • Nixon: 366 (520)
  • McGovern: 165 (17)
  • Schmitz: 1 (0)

Now that’s a striking difference. Nixon’s lead is cut down enormously, and John Schmitz of California, running as an American Independent, would’ve received 1 electoral vote from California. The result is still the same, and Nixon still wins by over 200 votes, but it would not have been anywhere near the landslide it was.

And we can see similar results with the 1984 election of Ronald Reagan vs. Walter Mondale, which is a larger landslide than Nixon’s re-election in 1972 and the largest victory margin since the 1788 and 1792 elections of George Washington. Again, applying the Nebraska model to the popular vote (actual in parentheses):

  • Reagan: 352 (525)
  • Mondale: 186 (13)

Again, very striking difference. Result is still the same with Reagan winning by nearly 2 to 1 in the electoral vote count, but that’s much more reflective of his actual popular vote margin of 58.8%, instead of winning under 60% of the popular vote but carrying almost 98% of the electoral college.

But clearly the case here is that the Nebraska model allows for third parties to pick up votes (provided the base votes are divided by overall popular vote instead of by congressional district), while also allowing both parties to pick up votes in most States. So it’s a much more fair breakdown in my opinion while still also preventing what could be a lot of elections being decided by the House of Representatives. It also diminishes the power of the largest States in the election.

Now again the numbers represented herein are presuming that the popular vote about proportionally represents how congressional districts would have voted. I’m aware that gerrymandering could affect this result. I’ll revisit this later once I have better numbers.