There has been some debate over whether the Democrat’s win was a “blue wave”. As Wikipedia states, “A wave election is a term from political science describing major gains made by a political party. Although there is no precise definition of what constitutes a wave election, the term is used in the United States when one party makes major gains in the House and Senate”.
I examined whether the current gains exceeded the average as well as the expected levels given the unemployment rate. I used data on the number of seats the party that held the presidency lost or gained in the midterm election from 1910 to the present (from Wikipedia). The analysis focused on the House as all the seats are competed every midterm, making these elections equivalent over time. In contrast, the Senate midterm elections involve different seats overtime with some years favoring one party or the other.
Beat the averages
First, some simple math. On average, the party that holds the Presidency typically looses 31.6 seats. When all the counts are final, President Trump is expected to lose 37 seats. In past midterm elections, the average unemployment rate was 5.9 percent, while it is currently 3.7 percent. Additionally, for most Presidents, the unemployment rate rises by 3.9 percent in the two years after election. Trump experienced a 4.2 percent decline in the unemployment rate. The positive unemployment situation should have made it easier for the Republicans.
Beat the prediction
Some more sophisticated math. I looked at how the unemployment rate and the change in the unemployment rate affected the change in the number of seats using regression. The data for these analyses are since 1950. Although using the unemployment rate to explain the change in the number of seats produced reasonable estimates, the model fit was poor. It explained very little of the variation in the number of seats won or lost.
In contrast, the change in the unemployment rate performed better. Although a very simple model (change in seats as a function of the change in the unemployment rate over the previous two years), it explains 15 percent of the variation in seats. The model predicts that if there was no change in unemployment, the incumbent President’s party would lose 23 seats. For every one percentage point decline in the unemployment rate, you gain 3.2 seats. Given Trump experienced a 4.1 percent improvement in the unemployment rate, the model predicts he should have lost 9.46 seats. Instead, he lost 37 seats. A blue wave.