Vinod B
Published in

Vinod B

Voting Trends and What it Means for the 2020 Election

When thinking about who will win the 2020 election, there are several stylized facts about voting behavior that can help us understand how people vote.

We can assemble data from the CCES (surveys run by a consortium of universities to study elections) and the Current Population Survey (run by the Census to find the employment rate) to examine voting patterns.

One of the stark patterns in voting behavior in the US is that geographic location is quite predictive of political party support. For example, below we plot the correlation in Democratic party support between states for US presidential elections from 1980 to 2016. We also group states together into their census regions. Red squares denote a positive correlation between states in while blue squares denote a negative correlation.

We see that within census regions the correlation is often highly positive, meaning that states in the same region tend to vote similarly in presidential elections. Also, across regions, some states vote quite differently from one another. The northeast and south often vote opposite as we would expect given the northeast tends to support Democrats while the South tends to support Republicans. Knowing the geographic location of a person then is highly predictive of how they are likely to vote.

Indeed, looking at the share of votes for Democrats in each state and election from 1980 to 2016 we see a remarkable consistency in many states over time. DC for example has voted strongly Democratic in every election while Wyoming has always voted Republican.

Another aspect of voting that this graph highlights is that states have become increasingly consistent in voting over time and more polarized. The number of states who switch their party affiliation between elections is lower than in the past and the degree to which states vote Democratic or Republican is increasing (as seen by the increasingly dark blue and red colors in the graph).

To see this in another way we can use something called PVI or Partisan Voting Index, which captures whether a state is more or less Democratic than the nation as a whole in a given election. Below we plot the PVI across states over time where lighter blue colors denote more recent PVI. We see that light blue dots tend to at either high or low extremes of PVI than darker blue dots, showing that over time states have become more polarized, which makes their expected voting behavior more predictable. It is incredibly unlikely today for example that California would ever vote Republican.

We see empirical evidence here for the increasing discourse around rising polarization of the US electorate.

Turning to demographics, there are also predictable patterns that are likely to repeat themselves in 2020.

Education has historically been highly predictive of party support. Below we show party affiliation by level of education across elections. We see that Democratic affiliation increases as education increases with the highest support amongst those with a 4-year or post graduate college degree.

Gender is also a dividing factor. Women are much more likely to be Democrats than men.

Likewise race is quite polarized where white people mostly support the Republican party but all other races are more more likely to be Democrats.

This intersects with education and gender as well where the typical Republican voters are now white men without a college degree.

Age also predicts party support with the probability of being a Democrat declining over time. Young voters are most likely to be Democrats while older voters are increasingly Republican.

Looking across the data we see that geographic location and demographics like education, gender, race, and age are highly predictive of voting behavior. Combined with the increasing polarization of the electorate overall and along these dimensions suggests that how individual people, states, and therefore the Electoral College turns out will be increasingly predictable mod the few traditional close states.

It is likely that we will see a close election again in 2020 as most states vote how they did in the past due to increasing sorting and polarization except for a few small states that will ultimately decide the election, raising the question again of why the US uses the Electoral College given its myriad problems.



Politics, Policy, Economics, & Data

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Vinod Bakthavachalam

I am interested in politics, economics, & policy. I work as a data scientist and am passionate about using technology to solve structural economic problems.