One of my favourite quotations is from H.L. Mencken:

There is always an easy solution to every human problem–neat, plausible, and wrong.

Tom Flanagan amply demonstrates the sort of reasoning at which this barb was aimed in a recent op-ed for The Globe & Mail, We don’t need a centre party to prevent polarization.

I will let his words speak for themselves:

What keeps democratic politics focused on the centre? Not the existence of a centre party but the workings of the “median voter theorem” (MVT). Think of voters as points spread out along a line – on the left, on the right, in the middle. By mathematical necessity, there is a median position, with half of voters to the left and half to the right. The median voter sits at the winning position in the democratic competition of political parties.

The proof is simple and elegant. If Party A moves to the left or right of the median, it allows Party B to locate itself closer to the majority of voters. The MVT predicts that Party A and Party B will tend to converge on the median because they cannot afford to let their rivals cut them off from more than half the voters.

The first problem is that he invokes the MVT as having some causal role here, as though it were a force moving people around rather than just a description of the phenomena. The MVT does not “keep democratic politics focused on the centre.” Any account that purported to explain such a thing would have to be vastly more complicated than this simple theorem. Its unsuitability to the task appears clearly when Flanagan notes that a move in one direction by a party “allows” the other part to locate itself closer to the majority of votes. The problem is that this “allows” not only is not “causes,” but is a stand-in for some entirely vague understanding of political strategy, and must admit of all sort of other determining factors which we don’t really have any idea of being able to outline without enough detail that we could consider the MVT to have real explanatory force here.

The other problem is that it is absurdly simplistic to lay out political view on a line. I thought first year undergrads learnt that any remotely sophisticated organization of the political spectrum does not draw its inspiration from a straight line. It’s surprising, and somewhat disappointing, that a political science professor such as Flanagan would give any credence to this approach.

This is not just the problem of relating abstract models to the real world. That is of course always a problem, as a model must abstract some things out in order to be a model and not just a copy. Flanagan correctly admits that “The MVT is a mathematical abstraction belonging to game theory, and the world is far more complicated than that,” but this makes it seem like the problem is just the traditional one of translating abstraction into real-world, concrete application. In fact the issue is that the model is just bad. And when you start with a bad abstraction, you will never get a good translation back into concrete terms.