The 1986 World Series was a close contest between the New York Mets and the Boston Red Sox that, unfortunate for Boston fans like myself, the Red Sox lost in 7 games. It featured some of the sport’s most iconic moments like Bill Buckner’s error.
Let’s imagine that series under a different method of scoring. Traditionally, baseball games are won by the team that scores the most runs overall, but what if each inning became an individual contest with the winner of a game decided by who wins the most innings? In this scenario the distribution of runs and when you score them matters, and it would have dramatically changed the 1986 World Series.
Prior to game six, the team that scored the most runs matched the team that won the majority of innings within each game, so nothing would have changed. In game six, however, things were different: through the first nine innings, Boston scored the most runs in three innings while the Mets scored the most runs in only two innings. Under this alternative scoring method, Boston would have won game six because they won three mini-inning contests compared to the Mets’ two, making them the winner of the World Series and changing the sports world forever.
No one argues that sports should be governed by this weird system of winners being teams that win the most innings, but this is precisely how we elect our Presidents. The Electoral College is a complicated, unintuitive system of individual, winner take-all contests at the state level (except in NE and ME) that is analogous to individual inning contests. It creates distorted outcomes where the winner of the Electoral College can differ from the winner of the national popular vote and dramatically affects the incentives of the political system, creating a bug in our democracy.
In two of the last five elections, Clinton vs. Trump in 2018 and Gore vs. Bush in 2000, we saw exactly the above scenario where the winner of the Electoral College and therefore the presidential election differed from the candidate who received the most votes, thwarting the majority of American voices. Those elections both shared the characteristic that they were relatively close contests with a popular vote margin within 3%.
This is not a recent accident: we see this in calculations based on a statistical model build to examine the likelihood of a clash between the Electoral College and popular vote winner using historical data on elections from 1980–2016 (when the Democratic and Republican parties formed their current iterations). The likelihood of a clash rises dramatically in close elections such that whenever the popular vote margin is within 3%, there is a one in three chance of the candidate who wins the majority of votes losing the Presidency because of the Electoral College. As the popular vote margin approaches zero, the chance of this happening becomes a coin flip.
And it isn’t just Democrats that are at risk of losing the Electoral College while winning the popular vote. In an average election, our simulations show that in close elections Republicans have about a one in four chance of losing the Electoral College while winning the popular vote. Indeed among recent elections we almost saw this in 2004 in Bush vs. Kerry.
Beyond just distorted outcomes that affect both parties, the Electoral College distorts the political environment by unevenly distributing voting power. It is typically argued that the system advantages small states, but this is not quite true. The Electoral College does reward small states on average more than large states, but what also affects the ability to swing the presidential election is how close a state is when voting. If a state is small but solidly Republican or Democrat, a voter has no ability to affect the outcome.
We can capture the balance of these two factors in the concept of Jersey Votes from the Princeton Election Consortium, which measures the likelihood of a voter switching the election outcome through their vote. Viewed through this lens voters in states like Pennsylvania have 200 times the influence of voters in New Jersey and over 2000 times the influence of voters in California.
Political campaigns are smart and recognize these discrepancies in voter influence, focusing their campaign dollars and visits to these states, ignoring voters in the majority of states and therefore ignoring the majority of the population. The uneven allocation of campaigning, spending, and voting power all combine to create huge differences in turnout such that battleground states see significantly higher rates as most voters feel powerless.
Lasting effects are not just limited to elections as political science research has shown that battleground states consistently receive more federal funds. Politicians are strategic actors who respond to incentives, and the Electoral College systematically creates incentives for elected representatives to skew policy towards the desires of these battleground states.
While geographic distributions of influence in the Electoral College often capture popular attention, they are only one aspect of voter identity, and indeed the current election system rewards certain demographics beyond location. Mormons, for example, are generally concentrated in states like Utah and Idaho, which are not particularly influential states under the Electoral College, making them one tenth as influential as the average US voter in determining the outcome of an election. Our current system effectively disenfranchises specific voters.
Given the huge inequities in the Electoral College and its downstream effects, it is time to reform. Critics argue that this is moot since Republicans have an institutional advantage under the Electoral College and will stifle any attempt at reform. While that is the state today, our work shows this has varied throughout history and with demographics changing in unpredictable ways in the future, it is unclear who will benefit tomorrow. It doesn’t make sense to have a system where voter influence is prone to the whims of demographic trends.
A popular vote also doesn’t mean Democrats will run away with elections. This is not how political parties operate. Instead, both parties will be forced to court more votes, necessarily tailoring their policies towards outcomes that the majority of voters support to remain competitive, decreasing polarization and pushing forward ideas that are popular among the general electorate.
The great thing about moving towards a popular vote is that reform doesn’t require every state to opt in. The national popular vote interstate compact is a state by state agreement to pledge a state’s electoral votes to the winner of the popular vote. Currently 11 states and the District of Columbia have signed on it. If enough states sign up to change incentives and make it unlikely for candidates to win the presidency without winning the popular vote, we can effectively make the campaigns act as if the national popular vote decided elections. Our calculations show that in some cases getting even single states like Florida to sign onto the compact can accomplish this, showing how close we currently are to changing things for the better.
We know that institutions matter because they create the system in which we exist and the rules that guide behavior. Having our presidential election system distort incentives of politicians in how they campaign, pass policy, and ultimately drive voters electing said candidates creates a system that is unable to effectively respond to the will of the people (not to mention the national security implications). It is time to change to another election system that better reflects equality of all voters.