Critics of the Electoral College often oppose the system because they see it as undemocratic, a system that enables a "tyranny of the minority," where a candidate with fewer votes ends up winning the election. The system's defenders then retort with appeals to the states. They remind us that "states are not mere subdivisions of the United States," as Justice Sandra Day O'Connor put it. The Electoral College, they explain, is necessary to make sure that the interests of all states, not just those with huge population centers, are considered by presidential candidates.
Upon first considering it, this argument seems valid. Small states get a boost in representation, so candidates can't ignore them. But in reality, candidates do ignore small states; we all know that. Presidential candidates focus their time and money on Florida, Ohio, and Pennsylvania, not North Dakota, Rhode Island, and Wyoming.
This is because the Electoral College has two salient features that distinguish it from a raw popular vote. First it gives some extra weight to voters in small states, but then it awards all votes to the winner of that state. Just looking at the population (favoring small states) or the closeness of the race (favoring swing states) will not give you an accurate picture of which states the Electoral College protects.
That's why I developed the Voter Importance Index, a system that measures the likelihood that any given voter in a particular state will determine electoral votes, adjusted for how many electoral votes represent that person. Put another way, the number of people needed to flip a state from one candidate to the other, divided by how many electoral votes those people vie for. This system provides a quantitative way to measure how much the Electoral College benefits voters in each state. You can read more about the index in the article linked above.
So, when it comes to the question of small states in the Electoral College, what does the evidence say? Analyzing election data from the past five presidential campaigns, the Voter Importance Index shows that the Electoral College does not, in fact, give any extra influence to voters in small states. In fact, there is a slight benefit to larger states.
Let's start with the 2016 election. Below is a scatter plot comparing states' populations versus their score on the Voter Importance Index. I use 2010 census population figures, because they determine electoral votes.
If the Electoral College gave extra power to small state voters, we would expect there to be a negative relationship between population and Voter Importance Score. As population goes up, importance goes down. But, as you can see, this is not the case.
As you can see, most of the states are clustered in the bottom left corner, with low populations and low voter importance scores. The trend line shows a very slight positive correlation between population and importance, but this trend is very week, only accounting for 0.1 percent of the variation in the data (r^2).
There are three outliers making this graph hard to read: California with its huge population, along with Michigan and New Hampshire, with their razor close elections. Taking them out of the sample provides an even stronger (but still weak) positive correlation.
The previous two charts only looked at data from the 2016 election. But the same conclusion holds up if we include data from the 2012, 2008, 2004, and 2000 elections, as shown below. The following charts still use the 2010 population figures because the number of electoral votes in each state does not dramatically change, and they use each state's average Voter Importance Score from all five elections.
At first, three outliers make it practically indiscernible:
When we remove the three outliers from this chart (California, Florida, and New Mexico) we find absolutely no relationship between state size and importance, a completely flat trend line, r^2=0.
The reason behind this pattern (or lack thereof) is really quite simple. The Electoral College doesn't benefit voters in small states because it doesn't really benefit voters in any state at all. If you look at the graphs in this article before we take out the outliers, you'll notice that almost all states are clustered along the bottom side, where voter importance is low. Only a handful of states gets substantial influence, and when they do, it is enormous. In 2000, for instance, voters in Florida were almost 3,000 times more likely to change an electoral vote than voters in Massachusetts.
Almost all states are worse off under the Electoral College. From 2000 to 2016, the mean Voter Importance Score for all states was 34.7, or 2,881 marginal voters for every electoral vote. This is a very high average, and it's being pulled up by states like Florida and New Hampshire, the real winners. Over the same time period, 44 out of 51 states and DC had an average score lower than 34.7. Voters in 44 states lose; voters in seven states win. You can see all this data on the spreadsheet linked below.
The only people who got extra influence from the Electoral College since 2000 are voters in Michigan, New Hampshire, Missouri, Iowa, Wisconsin, New Mexico, and Florida. Of the 15 states with the smallest population, only one (New Hampshire) is benefited by the Electoral College.
The data is in. We can stop blowing hot air. The Electoral College is good for swing states; that's for sure. But stop saying it protects the small states, because it just doesn't.
Photo credit D. Aderot, Creative Commons.
All Voter Importance Index data is available here.