If you are reading this, you might be one of the many Americans who are facing the prospect of voting for a lesser evil in this year’s presidential election. That prospect is painful. But there is nothing inherently wrong with making such a choice. Indeed, in many situations, it is likely the right thing to do.

Imagine an election where the only options are Queen Cersei from Game of Thrones, and Sauron, the Dark Lord from Tolkien’s Lord of the Rings. If Cersei wins, she will kill many innocent people, and oppress others. But she will leave much of the population more or less alone (as long as they don’t openly oppose her or threaten her family in any way). If Sauron wins, he will kill far more innocent people, and make the survivors his slaves.

You can instead cast a protest vote for a vastly better alternative, such as Gandalf or Daenerys Targaryen. But, by assumption, these are purely symbolic options, because they have zero chance of prevailing. If the protest voter would otherwise have backed Cersei, the net effect of his decision to protest is to increase the likelihood of the worst possible outcome: the triumph of Sauron.

Under those circumstances, it seems clear that a person who ensures a Cersei victory has done a good deed. He or she will have saved large numbers of people from slavery or death, even though the Cersei regime would be a deeply unjust one.

I. The Complicity Objection.

The most obvious objection to this line of reasoning is that you should not vote for Cersei because doing so makes you morally complicit in her evil actions. If you instead protest vote or stay home, you can remain untainted.

The complicity argument is intuitively plausible. But it is not as strong as it may seem. The voter in question is not responsible for creating the sad situation in which Cersei and Sauron are the only options. The net effect of his or her actions is a positive one: less death and slavery. And his intent is also good. He is not motivated by a desire to help Cersei commit atrocities. To the contrary, he abhors them, and is only voting for Cersei to avoid still greater evil. Sadly, the only way to do so is to ensure that Cersei wins. Whether you judge the voter’s decision by effects, intentions, or some combination of both, we must conclude that he did the right thing.

You can still reject this line of reasoning if you think it is never justifiable to back any evil at all, no matter how great a good is created by doing so. That’s a logically consistent worldview. But it requires adherents to bite a lot of bullets that few would actually accept. For example, it implies that everyone who backed the Allies during World War II was wrong to do so. After all, the allied governments (even the liberal democratic ones) were far from being paragons of virtue, and their triumph involved many injustices, such as the racist internment of Japanese-Americans. The same goes for supporting the American Revolution or the Union side in the Civil War, both of which were also far from morally pure causes. If supporting a lesser evil in war is sometimes defensible, surely the same applies to an election.

II. Does it Matter if your Vote Has Very Little Chance of Mattering?

A more subtle objection to voting for a lesser evil is the claim that you should not do so because the chance that your vote will make a difference is extremely low. Libertarian writer Jeffrey Singer explains the idea well:

1) According to Professor Ilya Somin in Democracy and Political Ignorance, my vote has, on average, a roughly 1 in 60 million chance of being the decisive vote in the Presidential election. (It might be a great as 1 in 10 million in my relatively small state of Arizona. It would have a roughly 1 in a billion chance of being decisive if I lived in California.)
2) If I vote for the lesser of evils and hold my nose, my vote is blended in with millions of others—there is no way to register my dissatisfaction with the choices the two major parties have given me. There is no way to separate those who voted for a lesser of two evils from those who voted because they actually LIKED the candidate.
3) If I vote for the Libertarian party candidate, I am directly affecting the vote total of that candidate. Because that candidate will get fewer total votes than the major party candidates, when all votes are totaled up, I will have had a greater effect on raising the total percentage of votes for the Libertarian candidate….

I appreciate Singer’s citation to my book, which is entirely accurate in so far as it goes. But he neglects another point I made in that same section of the book (and earlier in this article): While it is true that the chance of one vote being decisive is extraordinarily low, this is offset by the enormous effects if it does turn out to be decisive, after all. If, for example, the victory of the lesser evil means that each American will, on average, get an average income boost of $1000 relative to what he or she would get otherwise, that translates to $300 billion in extra wealth for society. And, obviously, the stakes are even higher if we are talking about lives and freedom, as well as money.

Even if you discount that figure by the infinitesmal likelihood of winning, it turns out to make logical sense to invest the small cost of voting for the lesser evil, at least if the difference between the two evils is at all substantial (the exact calculations are in the book). Sadly, it also means that it is often irrational for individuals to invest much time and effort in studying the issues at stake, which is why political ignorance is such a serious problem. Seeking out and studying political information requires a lot more time and effort than going to the polls. But that’s a different issue.

In addition, the low impact of voting weakens the case for casting a protest vote much more than it weakens the case for supporting the lesser evil. In a “first past the post” election where the candidate with the most votes win, there is always at least some small chance that one vote will be decisive. By contrast, when it comes to assessing the potential effects of a third party vote, there may not be any unique point where one additional vote clearly changes the way the protest is perceived. If the Libertarian candidate or the Green candidate gets 5.00000001% of the vote instead of 5%, will that really change anything? The chance that it will is surely far, far lower than 1 in 60 million. Perceptions of the meaning of third party votes depend on far more than just the vote totals themselves, which means that the potential significance of a protest vote is even more attenuated than a “regular” vote for a candidate with a serious chance of winning.

Finally, although Singer’s frustration with the two major parties is understandable, he is wrong to suggest that, other than protest voting, he has “no way to register my dissatisfaction with the choices the two major parties have given me.” To the contrary, there are many other ways to do so, including speaking out against the flaws of those parties’ agendas, as he himself has done.

Voting for the lesser evil on election day is entirely compatible with working to ensure that the available choices will be less evil the next time around. It is also compatible with advocating reforms that ensure that the power of government is more tightly limited, so that the next Cersei or Sauron we elect cannot cause as much harm. We must, as Conor Friedersdorf puts it, “tyrant-proof” the White House – and the other branches of government, as well.

I do not claim that this logic by itself tells us who we should vote for this November. Of the two major party choice, I think Hillary Clinton is a far lesser evil than Trump. But this year, unlike in most others, there is a real chance that a good third party candidate – Libertarian Gary Johnson – might be a serious contender. The odds are still against him, and we may well still end up having to choose between Clinton and Trump (or perhaps Cersei and Sauron…). But if it turns out Johnson has a real chance to win, the situation will be much different than in the usual, essentially binary, presidential election. As yet, it is too early to make a definitive assessment of his chances. I will likely have more to say about Johnson in a future post.

Be that as it may, there is nothing inherently objectionable in voting for a lesser evil. In some sad situations, it is likely to be the best option we have.

UPDATE: I recognize that the relevant calculations might differ depending on whether you are in a swing state or not. Swing-state voters have more of a chance of affecting the outcome than those elsewhere. But even a very small chance of your state being a swing state might still be enough to justify a “lesser evil” approach to voting, depending on how big the difference between the evils is.