Political science, prediction, and the US presidential election

The starting point for this presentation will be the fixation that some American political scientists have with predicting the outcome of presidential elections and the extent to which the developments in this year’s election are likely to confuse the predictive models even more than usual. This is not to suggest that the predictions are necessarily wrong, but rather that they throw little light on the most consequential aspects of the contest between Hillary Clinton and Donald Trump and that political science will need to shift its attention after the election is over towards more structural aspects of the US electoral process and to a more normative focus on party reform. This will necessitate revisiting a debate that took place some thirty years ago about the consequences of democratising political parties in the US.

In that context, the “success” of Donald Trump in winning the Republican Party nomination as its presidential candidate will be explained as a long-term consequence of the disastrous Democratic Party Convention in Chicago in 1968 and the subsequent transformation of the nomination process that began with the following election of 1972. It will be argued that Trump’s candidacy and, to a lesser extent, that of Bernie Sanders in the Democratic Party, illustrate some of the fundamental structural weaknesses of party reform in the US and that, irrespective of the outcome of this year’s election, both political parties will need to reassess their procedures for choosing presidential nominees. Political scientists can assist in that exercise if they can re-focus on the party reform debate of thirty years ago and tackle some of the underlying, but unresolved problems that were central to that debate.