Against strategic primary voting

I’ll vote for literally any Democrat in the general election, up to and including Satan himself or even Rahm Emanuel. But I just don’t see the benefit of “strategically” voting for a candidate I disagree with in the primary, due to some belief in “electability.” That is a purely speculative property. I don’t have the information necessary to decide that, and maybe no one does. Barack Obama sure seemed unelectable for a lot of common-sense reasons, but lo and behold, he actually got elected.

In any case I don’t trust the people who are trying to convince me of their personal theories of “electability.” Too much American political discourse takes place in those weird speculative meta-levels, where we’re supposed to choose the person we think other people will choose, or the person who will protect us from someone else. Every living American adult should be aware of the blackmail involved in the latter — and should be familiar with the disappointing results. Saving us from the worst looks an awful lot like the worst itself sometimes.

And again, we do not and cannot know for sure whether someone will actually win the election or protect us against the worst. What we do know for sure is each candidate’s policy proposals, and I think we should vote based on what we know instead of on our hunches about what other people (whose political preferences we don’t share or really understand) will think about the candidate at some future date.

The presidential candidate isn’t just a presidential candidate — they’re the leader of the party, who sets the agenda. Voting in the primary means voting on the direction of the party. If the Democratic Party is going to ask our opinion on that, we should give it to them sincerely, instead of psyching ourselves out through some ill-conceived 11-dimensional chess.

7 thoughts on “Against strategic primary voting

  1. In a (functionally) binary primary, I agree, but I could see an argument that, in a primary between Kucinich, Casey*, and Emmanuel, a lefty would be better off voting Casey over Kucinich in order to keep Emmanuel from the nomination. I agree that worrying about who can win the general is dumb; anyone who can win the nomination can, by definition, win 48% of the vote, and we can’t know/predict what happens with the next 2.1%.

    *that is, a fairly generic center-left Dem with better and worse positions

  2. Or you could assume that Emanuel supporters will be desperate to stop Kucinich and therefore voting for Casey, freeing you up to vote for Kucinich. It’s all game theory, and it all ends in infinite regress — but if they know that I know that they know I know….

  3. I agree with your main point, but you are over generalizing. Strategic voting can make sense in some circumstances.. ‘Don’t vote strategically in a presidential primary, in your own party, because of electability concerns.” Even “Don’t vote strategically in this presidential primary.” But voting strategically for posts other than president, or in a non-preferred party, or for reasons other than electability (or even over-electability) can make a lot of sense in some cases. As you say the presidential (and to some extent gubernatorial) votes effect not just who ultimately wins, but what direction a party takes whether they win or not. But that doesn’t apply in Senate, House, Treasurer, Statehouse, city council, and many other cases, and voting can be more strategic. Second, think of the case where you are happy with both leading candidates in your preferred party, but the other likely party has two leading candidates where you vastly prefer one to the other. In a open primary (or semi-open, or semi-closed), it can make sense to vote strategically (for the more preferred, not necessarily the less electable) choice in the opposite party primary. You are still using votes to express preferences not trying to game electability. OK third, plurality voting has a HUGE vote wasting problem. Strategic voting can be a way of minimizing that. Suppose your party has a primary candidate who is highly likely to win, say an incumbent governor popular in the party, with a minor challenger mostly trying to gain experience and exposure by running in the primary. But the opposite party has a more contested race (and isn’t hopelessly outclassed in the general). If you vote for the incumbent, your vote is likely to be wasted, because their victory is over-determined, and it doesn’t even send a helpful message to the party brass, who already know the incumbent is popular in the party. But if you vote for your preferred candidate in the opposite primary, it a strategic vote that genuinely expresses preferences, if not ultimate preferences, and isn’t necessarily a wasted vote. Blackmail is involved, but blackmail is at the heart of the mathematics of plurality voting. Only the supporters of the top 2 candidates get to vote for someone, everyone else only gets to vote against someone. (It can be unclear who the top 2 candidates will ultimately be, but this is still true, and terrible, and a great failure of plurality voting systems). Even in decent voting systems like instant run-off, or STV the strategic voting problem is not really eliminatable (see the Gibbard-Satterthwaite Theorem). But you are right that we should often resist the temptation to vote strategically, especially for presidential votes, or votes that strongly effect party direction, especially when outcomes are epistemically unclear, especially when we choose to vote for our preferred party primaries, and especially when our vote expresses guesses about outcomes but poorly embody our actual preferences …

  4. All good points. My argument is really against strategic voting based on “electability” in your own party’s primary. Even the kind of strategic voting that JRoth mentions may not fall entirely under my critique, now that I reflect further. Indeed, this may be a one-use-only “rule” that one should not vote for Hillary Clinton solely or primarily on “electability” grounds in the 2016 Democratic primary.

  5. What would (or, rather, what has) Žižek make of arguments from electability? (That is, I agree with this post, and I’m wondering if there’s not some psychoanalytic/marxian insight to be had here.)


    This link is to an article featured on Crooked Timber. It makes points counter to Kotsko’s. It also argues for Hillary instead of Sanders in general, regardless of voting strategy. It’s interesting this has been brought up, because I’ve recently been contemplating voting for Sanders. Like Kotsko said, we could drive ourselves nuts with the infinite regress problem and other complicated and unclear aspects of the process. Politics is… well, politics is politics.

    Someone told me they were “voting for Sanders for the primary, and Hillary for the Election”. Which of course assumes Hillary would win the primary, but on principle, they’d vote for Sanders because they found him most agreeable. And would vote for whichever democrat when the time came. We certainly have a responsibility to vote in our democracy, and that responsibility emcompasses considering the consequences of our votes for the election in November. Better Hillary than Trump? That’s easy to say, but we don’t know that it would necessarily come down to that or what the outcome would be.

    Is it compromising principles if one votes strategically, and not for their ideal candidate? Assuming voting strategically is not also voting as one would otherwise. Is it just playing the necessary game of politics for the least evil outcome? It seems this sort of debate occurs when we consider voting for third party candidates. The consequences are a bit different, the fear in the case of third parties is our vote becoming irrelevant if we vote “unstrategically”.

Comments are closed.