The Electoral College is elected on the first Tuesday after the first Monday in November (United States presidential election - Wikipedia).

The President is then chosen by the Electoral College no sooner than January 6 (Counting electoral votes in Congress).

Generally the members of a College publicly commit to a specific candidate in advance, so in practice the November election immediately determines the President-elect.

When was the last time, if ever, that the College ended up choosing someone in January other than the "winner" recognized immediately following the election in November?

NOTE: this question is related to Do any other countries take as long as the US to transfer government power following an election? - Politics Stack Exchange. Some answers are saying that the question isn't valid because the President isn't really known until January 6 at the latest. In theory that's true, but in cases where the November result is decisive, is it true in practice?

UPDATE: Based on the comments, what I'm asking appears to be unclear still.

It is not asking about situations where there is brief confusion at the time of the election (1948) about who won.

It is not asking about where some state results for the electoral college are contested.

In the typical case, one candidate concedes defeat and one accepts victory. In 2016 for instance, the vote was close, but during the night following the election:

at 2:50 a.m. Trump gave his victory speech. Later that day, Clinton asked her supporters to accept the result and hoped that Trump would be "a successful president for all Americans" — Wikipedia

Yes, it wouldn't be official until January, but for all practical purposes, Trump was considered the winner.

What I'm asking about is situations where everyone knows who "won", but some members of the EC later changed their commitment and as a result reversed the apparent victory.

  • 7
    Does 2000 count? Many networks incorrectly called Florida for Gore at the time, but they retracted those calls pretty quickly (and then called it for Bush, retracted that as well, and from there things got worse).
    – Kevin
    Sep 26, 2020 at 3:42
  • 3
    Do you mean the EC picked someone that was not who got the votes to get the EC majority, or where the apparent winner changed. What I am trying to get at is are you asking a thinly veiled question about one candidate appearing to win, but it changing when the mail in ballots are counted this year, or are you asking about the EC throwing a curveball by not all following their state election count. Also “announced on election night” is probably a relatively new (<100 years?) phenomenon.
    – Damila
    Sep 26, 2020 at 4:06
  • 2
    Note on your note: In modern times, 2000 was easily the latest that we “knew”. In the original days, Inauguration was in March.
    – Damila
    Sep 26, 2020 at 4:08
  • 5
    Does the famous 1948 "Dewey defeats Truman" newspaper headline count as an "announcement"?
    – dan04
    Sep 26, 2020 at 5:20
  • 2
    What, exactly, do you mean by “announced as [the] winner”? There is no such announcement, officially speaking, and of course anyone can unofficially announce anyone as the winner of an election.
    – Mike Scott
    Sep 26, 2020 at 5:52

1 Answer 1


The word you're looking for is Faithless Electors.

As far as I know, what you describe has never happened. To be clear, I mean that faithless electors changing the outcome of an election, not faithless electors in general.

According to Wikipedia, there has been only 165 faithless electors total in the history of the United States:

Over 58 elections, 165 electors have not cast their votes for president or vice president as prescribed by the legislature of the state they represented.3 Of those:

71 electors changed their votes because the candidate to whom they were pledged died before the electoral ballot (in 1872 and 1912). 1 elector chose to abstain from voting for any candidate (in 2000). 93 were changed typically by the elector's personal preference, although there have been some instances where the change may have been caused by an honest mistake.

Given the current size of the Electoral College (538), that is a very small number and already unlikely to make a difference. That certainly doesn't mean this couldn't have happened, it's just that at an average of 2 electors per election, that doesn't seem to have the numbers that would be necessary to do that.

Of course, better than speculation, it's better if someone were to know for sure. After doing some more research, I found a source seems to know for sure:

Faithless electors have never changed the outcome of a presidential election. To date, only one elector has cast a vote for the opposite party’s nominee instead of his own in a close contest. In the 1796 election - the very first contested presidential election - Samuel Miles, a Federalist elector from Pennsylvania, voted for Democratic-Republican candidate Thomas Jefferson instead of Federalist candidate John Adams.

So no, this has never happened.

  • 3
    Pulling some more specific numbers from that Wikipedia article: The greatest number of faithless electors in a modern (post-1900) election was 10, in 2016. This was not nearly enough to swing the election even if all of them had been defecting from the side with the most votes (i.e. Trump), but in fact most of them defected from Clinton. All other post-1900 examples listed in that Wikipedia article are single defections, except for the 1912 election, which had 8 (and IMHO isn't really all that "modern" anyway).
    – Kevin
    Sep 27, 2020 at 17:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .