In the first post of this two-part series, I explained how an NPV lobbyist admitted that their plan would allow state chief election officers (CEOs) to make up vote totals in the event that a state was unable to complete its vote count on time. This might happen due to a natural disaster, poor administration, litigation, or an intentional effort to interfere with the NPV compact. Regardless of what caused it, the result could be to take away or add hundreds of thousands or even millions of votes, potentially changing the outcome of a presidential election.
Although an NPV member state’s CEO could use their discretion to deliberately record zero votes or apply publicly available percentages to a state’s entire population, I doubt any would actually do so. More likely, the problem will be that there simply is not a single, obvious, universally-agreed upon method of determining vote totals for states that do not have official results available in time. As NPV’s lobbyists suggested, there’s a wide range of legitimate and understandable options, all of which produce different numbers and, potentially, different outcomes.
Suppose NPV was in effect in 2024 and there are problems in Kentucky, which has not joined the compact. Maybe there’s a recount, or legal challenge to the results or counting process, or a natural disaster (or given the way things have been going lately, all three at the same time) that delays the state’s counting of the votes and releasing final, certified totals, in time for NPV’s purposes (though the state might still be on track to meet its own and federal deadlines). As a result, twenty or so NPV member state CEOs are all trying to make good-faith estimates for the vote totals out of Kentucky. What are their options?
There are no statewide elections scheduled in Kentucky in 2024 other than for presidential electors, so simply applying the percentage – and I’ll cover that problem in a bit – to the vote totals from a gubernatorial or U.S. Senate race isn’t an option. There may be, however, one or more proposed constitutional amendments – there were two in 2020 – and it would make perfect sense to use the number of voters who cast votes on the constitutional amendment that had the most voters, which in 2020 was 1,825,749 votes.
But that would produce significantly lower vote totals than the real number, which in 2020 was 2,136,768 votes for all presidential candidates. So it would also make perfect sense to figure out the average difference between votes cast in past presidential races and on constitutional questions, and increase the vote totals being used by that percentage (just looking at 2020, there were about 17 percent more votes for president than for the constitutional amendment with the most total votes).
Those two different methods produce a spread of about 311,000 votes out of Kentucky. Both methods are legitimate, good-faith ways to estimate the vote count in a state that, for whatever reason, doesn’t have vote totals available on the timeline that NPV demands.
And there are other reasonable options that would produce different numbers. Using Congressional races, 2,115,895 votes were cast in 2020, only about 21,000 fewer than in the presidential race. But in 2016, there were only about 1,765,000 votes in Kentucky cast for Congressional races compared to about nine percent more, or just over 1,924,000, votes for president.
So in 2024, it would be perfectly reasonable for an NPV state CEO to just add a few thousand votes to the Congressional vote totals based on the 2020 results, and it would also be perfectly reasonable to average the 2020 and 2016 numbers and add about five percent to the 2024 Congressional numbers.
But wait, it gets worse. Because in this situation there isn’t an official percentage that’s been made available to apply to estimated vote totals. Instead, that same NPV state CEO that’s made a good faith estimate of the number of votes in Kentucky now has to try to figure out what the right percentage is to apply to those estimated vote totals.
Here too there are numerous reasonable options, any of which could be made in good faith. While the final percentages each candidate received wouldn’t be available, there likely would be partial results available that could produce percentages that could be used.
But that number based on partial returns may not be accurate, especially if the problem with Kentucky’s delayed election results are focused in just a handful of counties. Trump won Kentucky with about 62 percent of the vote in 2020, but Biden won Jefferson County (the most populous county in the state, home to Louisville) with over 60 percent of the vote. Assuming similar results for the 2024 race between the Republican and Democrat, if the problem is that Jefferson County hasn’t completed its vote count then the Republican’s statewide percentage would appear larger, and his or her vote total would be inflated by tens of thousands of non-existent votes while the Democratic candidate would have tens of thousands of very real votes removed.
Applying polling would be another defensible option, but they aren’t exactly precise. According to the site FiveThirtyEight.com, Trump was polling at around 55.6 percent while Biden was at 39.9 percent – in the end, Trump received more votes than polls indicated, while Biden received less. Again, tens of thousands of non-existent votes added for one candidate, tens of thousands of very real votes subtracted for the other.
Adding to the problem, not only are NPV member state CEOs expected to determine the number of votes cast for each candidate in a state where they don’t have good information, but there is no way to ensure that each states’ officials use the same method. The compact does not provide any mechanism for resolving differences or disputes, leaving each state to decide for itself on questions that could determine which candidate wins, or if a winner can even be declared. California might opt for one method while New York chooses another, while Massachusetts simply decides to record zero votes out of Kentucky. All would be permitted by the compact but would cause a political, legal, and constitutional crisis of epic proportions.
The television series Get Smart featured a bumbling international spy, Maxwell Smart, who would frequently be way off target on something, failing a task completely or forgetting some obvious or key fact. When his error was made evident, he’d reply with one of his stock lines, “Missed it by that much,” holding his fingers just an inch apart to suggest he was not, in fact, all that far off.
In one particularly appropriate episode, Max is told by his boss (“the Chief” – no other name was ever given) that the bags of cash nearly stolen by the bad guys were counterfeit. Examining it, Max observes that it seems to be real, it even has “a picture of President Goldwater on it.” The Chief reminds Max that “Goldwater lost the election by seventeen million votes.”*
“Oh yes, I remember now,” Smart replies. “Missed it by that much” (holding his fingers an inch apart).
Getting wrong who won the presidential election in 1964 was a funny bit on a funny show in the last half of the Sixties. Having compact member state CEOs explain to the country that they “Missed it by that much” because NPV forced them to concoct debatable and dubious estimated vote totals is a lot less funny.
*Goldwater actually received about 15.7 million fewer votes than Johnson.