Welcome back for Part II in what Trump's victory means for "democracy," whatever that poorly-defined term means. In Part I, we discussed biology! And a silly, little game called Ursuppe. The basic point was that you can't assess a system without sufficient data. So, let's get to the meat of the matter.
The core challenge for democracy is the following question: are ordinary citizens competent to make decisions, either about policy directly, or about who makes policy decisions?
James Madison, Alexander Hamilton and the rest were very skeptical about that, and we'll get to that soon enough. (If I were planning this out more carefully, I'd have done that first, but I write these posts during rather than after my morning coffee). Figuring out how often the public screws up, though, requires deciding how much data we have, and how much we have for alternative systems by way of comparison.
Now, linguistics diatribe. Data: plural. Datum: singular. Data are. Fuck you, Randall Munroe.
Anyway, how much data do we have on elections to assess citizen competence? Let's focus on the U.S. Shining light, beacon, blah, blah, blah. Do we count anything before women's suffrage? It probably wouldn't have changed any result, but in terms of legitimacy, yeah, kind of an issue. Then there's that whole race/slavery thing. Do we count pre-15th Amendment elections? Oh, and the 15th Amendment wasn't actually enforced before the 1965 Voting Rights Act, and even then, enforcement was a gradual process. (We could talk about Shelby County, but c'mon, we don't have lynchings anymore).
OK, so does that mean we don't get to count any presidential election before 1968 on the grounds that too many people were disenfranchised? That means we've got 12 presidential elections to evaluate. Small data set.
OK, so let's expand the data set. Let's throw in congressional elections, and other sub-national elections. Lots more data there! Here's the catch. People know less about those elections. What does that mean for democratic competence? We'll get to that in future posts!
Then the problem is about comparisons. To what do we compare our data set? Communist party systems? Easily managed, but a small data set too. Fortunately, those systems all sucked pretty hard, so we know where that comparison goes. Beyond that, there are so many odd variations of systems that have passed into the dustbin of history, many with poor record-keeping, that we have a different data problem. Suppose you want a complete, or at least random sample of princely states with mercantilist systems. How could you possibly generate your data set to evaluate the princes for their basic competence to see if they were fucking idiots or psychopaths on a Trump level or worse? See the problem? If we have a 1/12 chance of putting a Trump in power with our system, we outperform other systems if those systems put Trump-style nitwits in power with higher probabilities, but we can't know that unless we can generate data sets to figure out those chances, and we don't even have a good way of saying that our chances are 1 in 12 because that requires making a definitive statement about throwing out everything pre-1968.
This is a data problem. A messy one. Stay tuned, we've got plenty more, because Trump does nothing if not create messes for social scientists to address! (Well, he also grabs pussies...)