Democratic Chances Move Higher With President’s Approval Rating
https://www.economy.com/dismal/topics/election-model

The Moody’s Analytics election model has been updated to reflect the May economic forecast, which is relatively unchanged from the previous several months. Despite weaker than expected income growth, the forecast for a continued rebound in house prices as well as low gasoline prices portends a win for the incumbent Democrats. Unlike several other national models, our forecast relies on a two-year change in economic data as opposed to more high-frequency indicators.


The model predicts election results by state, assuming that whichever party wins the most votes in a state will receive all of that state’s Electoral College votes. (Two states, Maine and Nebraska, allow electoral votes to be split by congressional district.) The model’s dependent variable is the share of the vote received by the incumbent party in each state. Independent variables used to predict that share fall into three categories:

The two parties’ share of the popular vote in previous elections. This captures some of the noneconomic factors affecting election results, such as states’ general political leanings.
“Voter fatigue”—the tendency to turn out an incumbent party that has held the White House for two or more terms.
Changes in economic conditions before the election, specifically trends in gas prices, house prices, and real personal income.

 What The Model Says About the Election  Using historical economic and election data back to 1980, the 2016 model has correctly predicted the outcome of the last eight presidential contests. For the 2016 election, the Moody’s Analytics economic forecasts are used in place of historical data.