In this series, we are examining very close presidential elections in three states at three different times to understand how a very close election should be decided. In the first post, we looked at what happened in Pennsylvania in 1796, which we concluded should not serve as a model for the modern era. In this post, we’ll revisit the 1960 Kennedy-Nixon election in Hawaii. As we’ll explain, the way the recount was handled by all involved provides a model for how a very close election should be determined.
* * *
The 1960 presidential election between John F. Kennedy and Richard Nixon was held on November 8, 1960. That year, due to a slight difference in Congressional representation relative to today, 269 electoral votes (not 270) were needed for victory.
A day after the election, Kennedy was referred to as the “apparent” victor by the New York Times, but it was very close: a table on its front page showed John F. Kennedy having definitely secured 258 electors to Richard Nixon’s 172. Eight states with ninety-three electors remained formally uncalled, though the media thought that it was likely Kennedy would be able to secure enough votes to win. Hawaii was one of the eight states that was still uncalled.
The count in the close states continued even with Kennedy still the apparent victor. On Nov. 11, three days after the election, the Times reported Kennedy had just over a 0.5% lead in California. But that later vanished. Six days later, on Nov. 17, the Times ran a story titled “California Is Put in Nixon Column by Absentee Vote.” A red shift in the absentee vote had turned Nixon’s half percent deficit into a half percent surplus. (As it turned out, Kennedy did not need California’s 32 electoral votes that year, so that switch did not affect his prospects of becoming president.)
By Nov. 22, a full two weeks after the election, the outcome of the national election seemed somehow both certain and not. That day, the Times ran a story titled “Tally Is Finished in Only 10 States.” The article noted that Kennedy remained the winner based on “unofficial” returns, but that Republicans had also started to ask for recounts in a few close states. The Electoral College was scheduled to meet and vote around the country on December 19, so states had about a month to complete the counts before the appointed presidential electors were to vote. Time would be tight: the paper thought that thirteen states “may not complete the count until shortly before the Electoral College meets Dec. 19.” Illinois and Hawaii were among those last thirteen states to certify their official counts.
Illinois did not begin its recount until Nov. 29 and it did not conclude until Dec. 9. Kennedy remained the winner, by about 8,000 votes, and Kennedy electors were appointed in that state. (In 1985, a student of that election concluded that it had not been stolen by the Democrats.)
By that point, John F. Kennedy was not just the “apparent” victor but the presumptive president-elect with 300 electors, well over the 269 he needed. Only Hawaii with its three electors remained too close to call. And even though Hawaii’s votes did not determine the outcome, how all involved handled what happened next provides a way forward for us in the event of a very close election this year or any year.
Initial returns in Hawaii showed the Republican slate electors with a slim 141 vote margin over the Democratic slate. That was 0.08% out of 184,869 votes cast. On November 28, acting Governor James Kealoha, a Republican, signed a certificate of ascertainment appointing the slate of Nixon electors, whom he said had won by 141 votes.
But the Democrats still wanted a recount, and, as the calendar turned toward December, key deadlines approached for Hawaii officials. December 13, 1960 was the “safe harbor” day provided by 3 U.S.C. § 5. That means that, if an elector slate is certified as having been validly appointed on or before that day, Congress will presume that slate to have been validly appointed when the electoral votes are counted on January 6. At that moment in time, Governor Keahola’s November 28 certificate was the only one that could possibly get “safe harbor” status, so the Republicans argued in late November that a recount “would serve no useful purpose.”
State Circuit Court Judge Ronald Jamieson disagreed. Rather, on December 14, he ordered the recount to begin. Three days later, the Democratic slate edged ahead of the Republicans by twenty-one votes after thirty-four of the state’s 240 precincts had been counted. Kennedy’s lead grew to fifty-five by Sunday Dec. 18, the day before the presidential electors were scheduled to cast their electoral votes.
Let’s pause for a moment and take account of what each candidate had as of December 19, the day electors were required to vote under federal law. Nixon had in his favor the initial result along with the certificate, signed before the safe harbor date by the (acting) Governor of Hawaii. On the other side, Kennedy had in his favor a very slim lead in a recount that is not yet completed. So whose presidential electors should cast votes on December 19?
As it turns out—and this should serve as model for a close election this year or in any year—both sets of electors cast votes for their candidates. Yes, with the recount still in progress, two slates of electors gathered at the state capitol on Dec. 19 and cast competing sets of electoral votes. The Times confusingly reported that “Hawaii’s three votes went to neither candidate,” but clarified that what happened was “electors for both candidates voted for their man” and “Congress may have to decide” which votes to count.
Shortly afterwards, the Republican electors submitted their certificate of vote to Congress. Attached to it was the November 28 certificate of ascertainment certifying their appointment.
Meanwhile, on December 19, the presiding judge ordered the recount to proceed to count the votes in the entire state. The recount lasted for nearly two more weeks.
On Dec. 28, there was at last a winner, as declared by the Hawaii courts: it was Kennedy, by a nose. On that date, Judge Jamieson ruled that following a statewide recount the Kennedy slate had won by 105 votes. Two days later, on Friday, Dec. 30 he entered a judgment to that effect.
Kennedy had won. Yet only Nixon had the “official” appointment certificate, at least at that moment.
Fortunately, because both slates of electors had voted on the proper day, there was still a chance to tell Congress which slate was actually appointed by the voters. Indeed, correcting the record was required: a provision of federal law provides that “if there shall have been any final determination in a State in the manner provided for by law of a controversy or contest concerning the appointment of all or any of the electors of such State.” The state’s governor had to act quickly and transmit the revised result. Republican Governor William Quinn complied, and he composed and signed a revised Certificate of Ascertainment dated January 4, 1961.
On the morning of January 6, Congress received a package containing the electoral vote certificate composed and signed by the Democratic electors plus the revised Certificate of Ascertainment and Judge Jamieson’s judgment. (This was no mean feat at the time, given the distances involved and the fact that FedEx was still nine years from being launched.) Only hours later, “[t]he President of the Senate [would], in the presence of the Senate and House of Representatives, open all the certificates and the votes [would] then be counted[.]”
As the Senate’s presiding officer, Vice President Nixon opened the certificates from each state and handed them to the four Members of Congress serving as tellers who read, counted, and announced the results in each state. Following recent tradition, the states were taken in alphabetical order starting with Alabama. The process was unremarkable until it got to Hawaii. When they got to Hawaii, Vice President Nixon announced:
The Chair has received three certificates from persons claiming to be the duly appointed electors from the State of Hawaii. The Chair will hand these certificates one at a time to the tellers who will read the certificates and the attached papers in full. After the first certificate has been read the Chair will then hand the second certificate to the tellers, and then the third.
The first certificate was the package transmitted by the Republican electors. The second certificate was the package transmitted by the Democratic electors. The third certificate came from the Archivist of the United States to whom both packages had been sent, as required by law. The Archivist certified the authenticity of the very recently arrived Democratic package.
In perhaps the most gracious act of his long tenure in public office, Nixon said:
In order not to delay the further count of the electoral vote here, the Chair, without the intent of establishing a precedent, suggests that the electors named in the certificate of the Governor of Hawaii dated January 4, 1961, be considered as the lawful electors from the State of Hawaii.
If there be no objection in this joint convention, the Chair will instruct the tellers-and he now does-to count the votes of those electors named in the certificate of the Governor of Hawaii dated January 4, 1961-those votes having been cast for John F. Kennedy, of Massachusetts, for President and LYNDON B. JOHNSON, of Texas, for Vice President.
There was no objection. Congress had accepted a slate of electors certified just two days before the votes were counted. In the end, a slate of electors appointed based on a recount that did not begin at all until the safe harbor day and was not made statewide until the day the electors gave their votes.
That was the right outcome, though there is a sense in which the procedure feels disorderly. After all, when the Democratic electors voted on December 19, there was nothing official appointing them—there was only the possibility that their votes might be needed if the recount went their way. And then when Congress counted the electoral votes on January 6, it had before it competing certificates both of which, at a moment in time, carried the imprimatur of the Governor of Hawaii.
But in fact the dueling certificates, with the Governor later telling Congress who really won, was an excellent way to navigate a system that, for no good reason, occasionally provides too short a time to conduct a full recount in a very close election. The new President does not take office until January 20. Yet Congress requires that electoral votes be counted two weeks before then, on January 6. Congress also requires that the presidential electors vote approximately three weeks before that meeting, on the first Monday after the second Wednesday in December. The “safe harbor” date is six days before that vote. Looking at the calendar, then, states have only about a month to certify a count of popular votes before the safe harbor deadline.
Congress chose this timeline. It easily could provide more time for states to count votes and certify totals and somewhat less time to perform the mechanics of having presidential electors vote and then having Congress count those vote totals. Indeed, this year, Florida Republican Senator Marco Rubio introduced a bill that would set the safe harbor date as January 1, put the vote of the presidential electors on January 2, and maintain Congress’s counting of the vote on January 6. Perplexingly, this bill has gone nowhere.
So it looks like this election, like all those conducted since the current timeline was set in 1934 following ratification of the Twentieth Amendment, will allow only five weeks from election day to safe harbor day. That is likely not enough time to conduct a full recount in a very close statewide election. In addition to Hawaii 1960, a contemplated recount was not completed in Florida by December 12, 2000, when the Supreme Court ordered the recount halted and results certified as they stood so that the state could end the count by the safe harbor date that year. Other recent statewide recounts still pending as of five weeks after election day include the 2008 Minnesota Senate race, in which the recount was completed on January 6 (with Franken not seated in the Senate for six more months, until all appeals were exhausted), and the 2004 Washington Governor’s race, in which a winner was certified nearly two months after the election, on December 30, 2004.
Those races just let the recounts run until there was a winner, so there were not dueling claimants to office. But the race for presidential elector is different from the race for Senate in one important respect: although a state might not want to be without representation in the Senate if a result is not certain when the new term begins on January 3, a state would not lose that Senate seat for an entire cycle if a Senator is seated later than the first day of a new term. By contrast, electoral votes are use-them-or-lose-them: if, by elector voting day, a result is still uncertain and no presidential electors from a particular state cast votes, then Congress probably cannot count any electoral votes from that state for that particular election. The elector’s tenure in the office becomes defunct when the clock strikes midnight on Elector Day.
That means if a state wants to have its electoral votes counted, but which presidential electors were appointed by the voters on election day remains uncertain—because popular votes are still being counted—there is only one possible solution: both potentially-winning slates of electors should cast electoral votes on the day required while the recount continues. Then, if a recount is concluded by the day Congress counts the electoral votes, the state’s governor should—indeed, must—follow 3 USC § 6 and tell Congress which votes to count. Congress should then count the votes of the slate that won the legal contest. But if the recount is not complete when Congress counts electoral votes on January 6, then Congress must make the difficult call of judging which slate was lawfully appointed.
As strange as this procedure seems, it is accounted for in federal law. 3 U.S.C. § 5 gives safe harbor status only to slates appointed as a result of a “final determination” made six days before the electoral vote. Yet safe harbor status is not dispositive, and where a recount is ongoing, no slate will be able to claim this status. If there is a recount, 3 U.S.C. § 6 requires a governor to apprise Congress of who wins any formal election contest, even if that result comes after safe harbor day or even, as in Hawaii in 1960, after two slates have voted. Armed with that late-breaking information, Congress will be well-positioned and legally able to count the “correct” slate when it comes time to do so.
Yet while two slates of presidential electors could have voted in Florida in 2000 while the recount continued, that is not what happened. In the next post, we’ll explain why it could have and should have happened, and why the Supreme Court halted the recount instead.
* * *
 There were also fourteen “unpledged electors” from the Democratic south expected to be elected. If no candidate received a majority, those electors planned to extract a promise to oppose civil rights in exchange for their votes.
 We say “probably” because it’s legally unclear whether Congress could count votes of presidential electors cast on a day other than that set by law as elector voting day. Legally, the Constitution is explicit that Congress may set a date of voting for electors, and that day “shall be the same throughout the United States.” (This year, that day is December 14.) Votes given on a different date thus violate the Constitution and federal law, and are presumably invalid for that reason. The historical record is inconclusive from the one time we are sure a state’s electors did not vote on the required day. In 1856, a blizzard hit Madison, Wisconsin, which made it impossible for Wisconsin’s electors to meet on the required day. They cast their electoral votes the next day. Congress spent the better part of two days debating whether or not to accept the votes, but the debate ended without a conclusive resolution, since the state’s electoral votes would not swing the outcome. See Cong. Globe, 34th Cong., 3rd Sess., 644–60, 662–68 (1857).