« 100 Wins for Matt Cain | Main | Malvolio Production Diary: 9 days »

If you'd like to swing on a state

A quick election note: I recommend Pennsylvania Is Always Purple (And Other Electoral College Observations), a note by David A. Hopkins, a very acute (imao) political scientist who blogs at Honest Graft. His stuff is generally worth reading, actually.

I bring up this piece because it delineates some of the odd things about our time in American Politics. It’s a sort of paradox—on the one hand, swing states or battleground states or whatever you call Florida, Ohio, Pennsylvania, North Carolina and Virgina (and maybe some others) don’t actually swing independently of each other or the nation at large. If you figure there is some more-or-less true national swing, some number N such that Clinton’s share of the vote minus N equals Trump’s, then all states, including swing states, have a number k, such that for each state S, Clinton minus N plus k equals Trump. I should say that N and k can be positive or negative, and I should probably stop speaking in algebra altogether… the point is that Clinton will do seven points better in Connecticut than she does in the national popular vote, and twelve points worse in Alabama. And that k, that difference from the national average, is smallest in these swing states, largely due to the demographic makeup of those states. Broadly speaking, if Clinton gets 70% of the Latino vote, she will get 70% of the Latino vote everywhere, and the numbers in a state will be more about how many Latino votes are available than how well she does among Latinos in that state.

Did you follow any of that? Well, in short: there are swing states, but they follow the national trends just like everywhere else. They are just balanced close enough to 50% that the plus-or-minus of a candidate can sway the outcome one way or another, while in the others, the national difference is not going to be enough to push the state into the other column. Clinton would take Alabama if she won the popular vote by fifteen points, but it isn’t going to happen. So from the standpoint of an interested observer, you are better off looking at national polls than state polls; they will tell you more about the eventual electoral college outcome.

Unless you are deciding where to put resources for a campaign. The campaign should put its ad money in the swing states of course… even though the swing states, as I just said, don’t actually swing independently. So what does all that money actually do?

Well, who the hell knows? I believe that ads aren’t actually very persuasive, but you don’t need to believe that to believe that two campaigns running a lot of ads cancel each other out. And of course even if a superior ad campaign and GOTV and the rest of it move the outcome by only a tenth of a percent, that may well turn out to be the tenth of a percent that matters, if the election is close enough.

The paradox is that campaigns must spend the bulk of their time moving the parts that move least, by pushing on them in the spot with the least leverage. They don’t have any choice, of course. Presidential elections are fought locally but won nationally. And I don’t even believe that they are won or lost nationally in the campaign, in the usual run of things. They are won or lost by the incumbent party governing well or poorly, in the eyes of the national electorate. Which is as it should be.

All of which is the lead-up to saying this: if you are considering donating money to a presidential campaign, please don’t. Find one of your local races, or a Congressional race somewhere in the country. Those races are won locally, and the money might even do some good.

Tolerabimus quod tolerare debemus,

Post a comment

Please join in. Comments on older posts will be held for moderation. Don't be a jerk. Eat fruit.