Three variables for 2018

This post was originally published on this site

Ten days ago, Alan Abramowitz, an elections expert and prominent political scientist, posted an essay showing how three variables can predict with considerable precision the outcome of midterm congressional elections.  The party that controls the White House, the number of seats held by each party prior to the election, and the difference in the “generic ballot test” together explain “90% of the variation in seat swing” in midterm elections from 1946-2014.

Prior to the 2016 presidential election, Abramowitz wrote about how just three variables can also be used to predict the outcome of presidential elections.   Despite all the controversy, scandal, and uniqueness of the 2016 election, knowing the incumbent party’s status in the White House, presidential approval in June, and second quarter GDP growth (the “fundamentals” or “time for change” model), a person could have predicted within 2 percentage points the outcome of the two-party vote.

Turning to the 2018 race for governor in Iowa, previous research has shown that three variables matter when it comes to predicting gubernatorial approval: “relative unemployment” (the difference between federal and state unemployment), presidential approval (shared party affiliation with a(n) popular(unpopular) president), and time spent in office.  This is important because approval ratings are generally good indicators of what will happen on Election Day.

My own research on Iowa governors added a fourth variable: the level of “comfort” Iowans have toward the chief executive.

There will be a lot of twists and turns along the way to November 2018, but it is important to remember that a few fundamentals can go a long way toward predicting election outcomes.

If you are looking for something to watch in 2018, keep tabs on the Iowa economy, presidential approval, and the sense of connection Iowans have toward candidates on the ballot.

(Visited 16 times, 1 visits today)