The difficulty of estimating the effects of "voter suppression" techniques

There are a lot of tricky policies at the state level that Republicans have been using for the last few years out of the ostensible fear of "voter fraud." Voter fraud, of course, doesn't exist. Justin Levitt, at Loyola, has tracked every single accusation of voter fraud ever made, and it just doesn't happen. Does that mean the people imposing policies to respond to it don't believe their own bullshit? No. People can convince themselves of anything, if they have partisan incentives to believe it. The term is motivated reasoning. However, at least some of them know it's bullshit, and that they are just suppressing votes from those likely to vote Democratic. Estimating the effects of these policies, though, is very difficult.

I'm going to make a simple observation this morning about the general trend rather than any specific policy. Remember the polls in 2016, and how they sucked? Here's the thing about voter suppression techniques. If they work, they make the polls suck. A group of people want to vote, and can't. That biases the polls in the direction of the Democrats, such that the GOP overperforms on election day. Sounds familiar, right? Here's the catch. The extent to which any state attempts to suppress non-white votes varies at the state level. (Let's not mince words. This has a big racial component.)

So, which states had polls that were the most wrong? Those would be the rust belt states. (Hi!) Those weren't the states with the most aggressive policies! That suggests the effects at least aren't what you might think. Calculating those effects is hard. That doesn't mean they don't exist, just that they are small. Why? Most policies either work on people who have low turnout rates anyway, or can be worked around with organization for the subgroups with a higher propensity to vote. Voter ID laws can be worked around with... IDs. (Except in North Dakota with the Reservation thing, and Heitkamp is toast anyway). Purged voter rolls can be worked around by filing provisional ballots, generally. There are workarounds for many, although not all voter suppression techniques. What does that mean in the aggregate? It means that the effects might not be as big as some expect.

So finding the effects is hard. Can they matter? Sure. In a close election. But, in a close election, anything can matter. In 2018, states under GOP control are trying everything. Will their attempts matter? The evidence so far is not that strong, and it depends not just on what they try, but where they try it.