The Fair model on 2008

During election season, there’s a lot of talk about “keys to the election,” all the factors that supposedly influence the outcome of the presidential election: the economy, Iraq, the candidates’ charisma, the campaigns’ organization, fundraising, media bias, etc. But Yale economist Ray Fair would argue that there is exactly one key to the election, which is the economy.

More precisely, he shows that you can predict the outcome of presidential elections using an equation, the Fair model, that is based on economic variables and a few other predictors (e.g., incumbency and party). Notably absent from the equation is anything referring to current events, or to the candidates themselves. According to Fair, you can predict the outcome of the election as soon as you can predict the economic variables, long before you even know who the candidates are.

The equation has successfully predicted the popular vote of each presidential election since 1996, and retroactively predicts nearly every election back to 1892. (An earlier version of the model, developed in 1978, was revised after it failed to predict the outcome of the 1992 election.) In 2000 it essentially predicted a tie, with a microscopic margin dwarfed by the margin of error.

For the 2008 election, it predicted that Obama would win 51.9% of the two-party popular vote, with a standard error of 2.5%. The actual outcome was 53% for Obama, meaning that he slightly outperformed the prediction, but was well within its margin of error. The striking fact is that Fair made similar predictions as far back as November 2006, long before we had any inkling who the candidates and what the issues would be.

Given this, it is hard to escape the impression that most presidential elections are largely a charade, a lengthy process toward a pre-ordained outcome. All the “keys to the election,” including the candidates themselves and all the issues, amounted to just a 1.1% deviation from the prediction. An optimally Republican-favoring campaign season, one that swung the result by the entire standard error, would have eked out a 50.4% victory in the popular vote. (Recalling 2000, such a narrow edge might have been insufficient for an electoral college victory.) In any case, this year was hardly optimally Republican.

I find these results somewhat dispiriting, because it makes no sense for voters to make their decision solely on economic variables. The President has relatively little power to affect the economy, compared to his much greater power and importance in foreign affairs. And frankly, foreign affairs are much more important, particularly today. But alas, according to the Fair model, foreign affairs hardly figure in to the election at all.

Another takeaway from Fair’s results is the silliness of the idea of an electoral “mandate” for the President. Since you can predict the election’s outcome without knowing the candidates, much less any promises the candidates have made, the election can hardly be viewed as a judgement on those promises.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s