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?


Don’t Mess With Texas (Recounts)
Sean Parnell • Mar 25, 2020

One significant danger of the National Popular Vote Interstate Compact (NPV) is that it would require a nationwide recount in the event of a close election. This would be a nightmare. For what it’s worth, that’s not just the assessment of those opposed to NPV, it was stated by a lobbyist for NPV at a hearing I recently attended.

NPV advocates’ general response about recounts is, and I hope I’m fairly describing it, that recounts are pretty rare – I believe I’ve heard it stated by them that only one in every 200 presidential races under NPV might require a recount – and that state laws would apply, so basically there’s no reason to worry.

That seems unsatisfactory, at least to me. The opportunities for chaos and confusion under a recount are endless; I’ll outline a number of them in future posts. But right now I’d like to focus on what might have happened in Texas had NPV been in effect during 2000, when then-Governor George W. Bush won the election despite then-Vice President Al Gore receiving more votes nationally (i.e. he was the “national popular vote winner”).

The Texas Secretary of State’s office has a helpful web page on the subject of recounts. It explains that only the losing candidate can request a recount, and there are only limited grounds for filing a recount request. It describes the most relevant as follows:

Difference between the number of votes received by petitioner and number of votes received by the person who was elected or is entitled to a place on the runoff election ballot is less than 10% of the number of votes received by the person elected or entitled to a place on the runoff ballot (same formula for votes for and against a measure).


Jane Doe 2000

John Doe 1850

difference 150

10% of 2000 is 200, therefore, John may request a recount

In 2000, the national margin between Bush and Gore was about 547,000 popular votes, out of more than 100 million votes cast nationally. Clearly, the difference between the candidates is less than 10 percent of the winner’s total votes.

But the vote in Texas wasn’t nearly as close – 3.8 million for Bush to 2.4 million for Gore – well above 10 percent (about 37 percent).

Question: Who would have been the “loser” in Texas in 2000 if NPV was then in effect, and thus who would be able to request a recount?

  • Bush, because he has fewer votes nationally than Gore?
  • Gore, because he has fewer votes in Texas than Bush?

Question: Was the difference between the winning candidate’s total votes and the losing candidate’s votes less than 10 percent of the winning candidate’s vote total?

  • Yes, because the margin nationally is less than 10 percent?
  • No, because the margin in Texas is more than 10 percent?

These conflicting determinations would ultimately, of course, have to be settled in court (which raises another question – state or federal court? The answer is, again, unclear and could really be either). It’s simply unknowable how those judges would rule, and also almost certain that those rulings would be appealed. The losing side would most likely cry foul, see political bias, and demand another rules change.

These sorts of questions – and there are literally thousands more just like it, based on the specifics of any given contest and the election laws of the 50 states plus Washington, DC – have no obvious answers. The NPV compact fails to say anything about resolving disputes or recounts. A close national election under NPV would cause chaos and partisan rancor on a scale that nobody sane wants to endure. And given that three of the past fifteen presidential elections have had popular vote margins of between about 113,000 and 547,000, it seems like NPV’s prediction of only one in every two hundred elections being catastrophic might need to be revisited.