Blog

One of the many defects in the National Popular Vote interstate compact is that it would be essentially impossible to conduct a full recount* if the national margin were close. This is because state recount laws are all different. Instead, individual states would decide whether the close national margin (or perhaps a close in-state margin) would either require or allow a recount. This would result in a partial recount, with various states operating according to their own rules—and their political biases.

And here’s the fun part: it’s impossible to know which states will decide to recount the votes, and which won’t. There isn’t a single state with a recount law specifying what to do in the event of a nationally-close presidential election. Instead, election officials and (ultimately) state and federal judges will look at individual state laws and regulations to determine what is required or allowed under state laws. All those laws were written assuming a recount might be necessary if an in-state (or in-district) margin was close, but none ever contemplated a “national popular vote.”

Consider the Illinois recount law, which allows a losing candidate to request a recount if his or her initial vote total is 95 percent or more of the winning candidate’s total (there are no automatic or required recounts). So in a statewide race, such as for governor, if the candidate initially determined to have won received two million votes, then the second-place candidate would have to have received at least 1.9 million votes in order to request a recount.

There are two very large problems for NPV if there’s a close national election and one (or both) candidates are trying to get states to recount the vote. Suppose the national margin is 0.1 percent (roughly 150,000 votes) in favor of the Republican ticket, while the Democrats won Illinois 58 percent to 42 percent (about what it was in 2020). The Democratic campaign believes, reasonably enough, that a recount in Illinois would likely turn up thousands or even tens of thousands of additional votes that favor them, based on the state’s partisan tilt and the fact that there’s more than 100,000 undervotes, overvotes, and rejected absentee and provisional ballots to go through (about what it was in 2020).

The first problem, of course, is that Illinois’ law was written with the in-state margin in mind, and the losing candidate did not receive 95 percent of the votes cast for the winning candidates. Does the state ignore the in-state margin and instead rely on the national margin to determine if the state conducts a recount? Maybe, but maybe not—this would likely be up to an Illinois state judge, although a federal judge could try to step in.

The second problem is that it’s not clear who are actually the “losing” candidates! The Democrat ticket in this scenario received the most votes in Illinois, so can they really be considered the “loser” under Illinois law? Does the state again rely on the national results and determine the Democrats are the “losing” candidates in this scenario, allowing them to request a recount? Again, maybe, maybe not—judges will decide.

It’s easy to think that Illinois courts might go ahead and help the Democrats out here, ruling that since the state is in the compact, it’s the national margin that is relevant, so the Democrats can request a recount in Illinois. So, problem solved? Not exactly, because of course there are 49 other states (plus Washington, DC), all of which have recount laws of their own (except perhaps Mississippi), not all of which are going to arrive at the same decision as Illinois and allow a recount to proceed. 

Consider Texas, which has a recount law similar to Illinois—there must be less than a 10 percent difference between the winning and losing candidate, and only the losing candidate may request a recount. Suppose that in that same election, the Republicans—who won Texas 52-46 percent (roughly the 2020 margin)—thought a recount in Texas would help them, potentially offsetting any new votes the Democrat found in Illinois and elsewhere.

There are two probably insurmountable problems though, unlike Illinois. Texas is not in the compact, so its courts cannot argue that compact membership requires them to look at the national rather than the in-state margin (a 52-46 margin would fall just outside of the required threshold). That doesn’t make it impossible for a court to wave some judicial theory around and claim that since the national margin is what counts, they’re going to use it instead of the in-state margin. But it does make it a lot harder and less likely.

The second problem here is that, unlike in Illinois, there’s no real question about who is the winning and who is the losing candidate—the Republican is the winner in Texas, and the Republican is the winner nationally (at least before any recounts). And Texas’ law specifies that only the losing candidate may request a recount.

In this scenario, what happens is that Illinois conducts a recount, probably one that is turning up thousands or even tens of thousands of new votes for the Democrat, while Texas does nothing. And every other state will similarly have to figure out how to apply their recount laws, resulting in some states recounting while others don’t, and some states adding into the count absentee and provisional ballots that were initially rejected while others only recount the same ballots that were counted the first time around. 

In the end, it’s the state-by-state “patchwork” that the U.S. Supreme Court feared in a recent decision concerning presidential elections (suggesting yet another reason the compact is vulnerable to being struck down). Which is sure to drive partisan rage, particularly if most of the states conducting recounts overwhelmingly favored one candidate while the non-participating states favored the other. I can’t imagine Republicans being happy if most “blue states” are recounting while most “red states” aren’t, nor can I see Democrats being happy in a scenario where many “red states” are counting but most “blue states” aren’t. It certainly wouldn’t be fair, and seems like a recipe for disaster. It’s yet another reason to reject NPV.

* Because I didn’t want to write a 7,000 word post, for simplicity I’m lumping election “recounts” and “contests” together and gliding over some of the nuances and distinctions between the two. Generally recounts occur before results are certified while contests occur after certification, if you want to learn more, try Ballotpedia.