Now that the presidential election is over there’s a lot of reflection and analysis of the electoral and popular vote. Some of the presidential campaign disagreements centered around income inequality yet I fail to see any analysis of the impact of existing income inequality on the popular vote.
Partly for fun and also to determine the empirical impact of income inequality on the election outcome I found the popular vote outcomes by county at The Guardian and the latest county Gini coefficients, measuring growing inequality from 0 to 1, at the American Community Survey. I then created a new binary variable for an Obama win in each county (1 = win, 0 = loss). As the dependent variable, Obama_win, is a binary variable, logistic regression is an appropriate statistic tool of choice.
The logistic regression revealed the predictor variable, income inequality measured by the Gini coefficient, was statistically significant (p > z = 0.000). Utilizing the regression outcome I then predicted the election outcome in each county and then plotted the prediction against each corresponding measure of inequality, the Gini coefficient. The following graph summarizes the results.
It’s observed as income inequality increases, President Obama’s probability of winning a county victory increases.
The GOP is examining their positions on Latinos, women and a number of other issues. Perhaps they should consider the adverse effects of income inequality on their probabilities of winning votes.