Your web browser is out of date. Update your browser for more security, speed and the best experience on this site.

Update your browser
Defending the Electoral College and the Constitution since 2009

what are you looking for?


NPV still has an RCV problem
Sean Parnell • Aug 20, 2021

One issue recently generating concern about the National Popular Vote interstate compact (NPV) is its fundamental and profound incompatibility with the election process known as Ranked Choice Voting (RCV). Attempting to determine a national vote count under NPV would encounter serious problems if any state used RCV in its presidential election.

Under RCV, voters rank candidates in order of preference. If one candidate gets a majority of first-place votes then they win. If no candidate receives a majority then the lowest-ranked candidate is eliminated and those votes are transferred to those voters’ second choice. The process continues until one candidate receives a majority.

RCV poses at least three major problems for NPV, in particular if the national margin is close (as it was in four of the last sixteen elections).

  1. States using RCV can decide to report either initial or final (RCV-adjusted) vote totals, and candidates could gain or lose hundreds of thousands of votes depending on which vote total is reported.
  2. States using RCV can decide to report both numbers, leaving it to NPV compact member state officials to decide which vote totals to use.
  3. If a major party candidate finishes in third place in any state using RCV, hundreds of thousands or even millions of votes could be completely erased from that candidate’s national vote count.

At best any of these scenarios would spark a national uproar and lead to a flood of litigation over which vote totals are “real” and should be used. At worst, it would be a political, legal, and constitutional crisis of epic proportions.

Advocates for NPV only recently acknowledged the problem that RCV poses for NPV. FairVote, an organization that supports NPV but appears focused primarily on promoting RCV, wrote last week in a blog post that legislation recently passed in Maine addresses the problem and ensures that RCV “works well with the proposed National Popular Vote Interstate Compact.”

Unfortunately, it doesn’t.

The legislation in Maine dealt with a number of election issues, including a requirement that the state’s “Certificate of Ascertainment” include only the final, post-RCV process vote total for each candidate (the legislation is here, the relevant section is at the end).

While the legislation would seem to address the first two problems, it does not really fix either of them and may make the third problem worse.

  1. The law directs Maine’s governor to report post-RCV numbers on the Certificate of Ascertainment, but that is not the only document that official vote totals can be obtained from for NPV’s use. As compact advocates have pointed out repeatedly when I have documented problems with these certificates, states typically produce a separate certified statewide canvas of votes or similar document. Would that document, which can also be used under NPV to obtain vote totals, include both the initial and final vote counts? The answer is likely yes – the official results on the Maine Secretary of State’s office for the 2020 Republican primary in the 2ndCongressional District include both initial and final numbers, and there’s no reason to believe the state wouldn’t or couldn’t as part of its statewide canvas for a presidential election include that same information. There’s no requirement that officials in NPV member states use Maine’s Certificate of Ascertainment as its source for vote totals, particularly if there’s a statewide canvas available as well. Both the initial and final vote counts are likely to remain available on official election results available for use by NPV states.
  2. Even if the Maine legislation resolved the issue of which set of vote totals should be reported, it only addresses Maine – other states that use RCV (Alaska will use it in 2024, and other states may adopt it as well) would pose the same problem for NPV. It seems unlikely that states hostile to NPV would agree to adopt legislation aimed at helping other states operate the compact – in fact, several states have considered legislation explicitly designed to thwart the compact.
  3. The Maine legislation literally locks in the third problem, guaranteeing that a major party candidate who comes in third behind an independent or third-party candidate will have hundreds of thousands or even millions of votes completely erased from their national vote total. This is not farfetched: Maine is one of two states (Utah is the other) where a major-party presidential candidate finished behind Ross Perot in 1992 (Perot narrowly beat George H.W. Bush in Maine and Bill Clinton in Utah). In that election, almost 400,000 votes would have been erased from national totals if RCV had been in effect in those two states. The Maine law doesn’t just lock in this problem, it makes it more likely to occur because of how the state runs its RCV process. As the FairVote post explains, “When there is no first-round winner, Maine will run its final statewide tally down to two presidential candidates and put those final-round vote totals on the certificate of ascertainment.” This means that even if one candidate receives a majority of votes before the third-place major party candidate is eliminated, the process will continue until that third-place candidate has every vote erased. (This would quite possibly violate the constitutional rights of many voters.)

To illustrate how this would occur, consider a scenario similar to what happened in Utah in 2016. Donald Trump easily won the state with 45.5 percent of the vote, Hillary Clinton finished in second with 27.5 percent, and Evan McMullin received 21.5 percent. Other candidates, nearly all of them right-of-center, received the remaining 5.5 percent.

Now suppose Utah used RCV the same way Maine intends to, and both Trump and McMullin each did two percentage points better than they actually did while Clinton did four points worse. Under this scenario, initial results are Trump at 47.5, Clinton with 25.5, and McMullin gets 23.5. If the 5.5 percent for the other right-of-center candidates split evenly between Trump and McMullin, then after their votes are re-allocated Trump would have a majority (50.25 percent), McMullin would finish second (25.75 percent) and Clinton would be pushed into third.

Under some RCV systems, the RCV process would now be over – Trump received a majority, and there is no need to continue the RCV process. Hillary Clinton’s roughly 288,500 votes from Utah would remain part of her national vote count. But if Utah runs the RCV process until there are only two candidates left, as Maine intends to do, those 288,500 Clinton votes are erased, at least from Utah’s Certificate of Ascertainment.

Erasing 288,500 votes for a major party candidate wouldn’t matter much in a race where the margin between the top candidates was several million votes. But in a close election, such as those in 1960 (historians still argue whether Kennedy won the popular vote by 113,000 or Nixon did by about 60,000), 1968, 1976, or 2000, having several hundred thousand votes disappear could make a difference, especially if it happened in multiple states.

Of course, as mentioned previously those 288,500 votes would probably still exist on the Utah certified canvas of results, and it isn’t hard to predict in a close election a lot of pressure in the NPV member states to go with the “real” vote totals that don’t erase a massive number of votes. The compact doesn’t provide any guidance, and each compact member state would be free to go its own way – some might use the “top two” numbers on the Certificate of Ascertainment, others could opt for the statewide canvass with all the votes.

FairVote apparently recognizes that these problems still exist – the post states that they “envision[] further reforms that could harmonize the NPVIC with ranked-choice voting to ensure majority rule and political equality. FairVote’s Rob Richie and Jeremy Seitz-Brown, along with Patrick Hynds, Stevie DeGroff, and David O'Brien are coauthors of a forthcoming Harvard Law and Policy Review article that details potential federal legislation and a potential additional interstate compact among states using ranked-choice voting in presidential elections.”

Federal legislation seems an unlikely solution – while we at Save Our States have our differences with the folks at National Popular Vote about precisely how broad the power of states is to allocate their electors, we certainly would agree it is broad and probably precludes Congress from passing a law requiring compact member states to calculate the national vote total in any particular way.

Regarding an additional compact for RCV states, this seems to still leave the problem of officials in NPV member states using vote totals from the official statewide canvass instead of the numbers on the Certificate of Ascertainment. So it looks like there would need to be a third interstate compact to iron these problems out, this one for NPV member states dictating which results to use and how to use them. (And then what if a state adopts Approval Voting?)

Ultimately, the combination of NPV and RCV creates problems with no acceptable solution. Even if the suggested additional compacts were to go into effect with NPV, the likelihood of erasing hundreds of thousands or even millions of votes from the national count would remain. It’s hard to see this being acceptable to anyone.

Image Credit: (AP Photo/Mary Altaffer)