Do You Trust the Government with the Power to Change Your Vote?

The 2016 election is now well past and it’s a good opportunity to sober look back on the electoral process. There was a great deal of partisan commentary about the electoral system, and very little of it made any sense. Further, I think that if the roles were reversed, and Clinton won the electoral college without even a plurality of the popular vote, the commentary would also be reversed. Supporters of Clinton would remind us that we don’t live in a democracy but rather a constitutional republic and demand that any Trump supporters protesting the election results stop whining and go home.

Donald Trump is the president, so any lesser of two evils type argument no longer applies to either of the former candidates and we can wonder why we just had an election where the only two viable candidates were historically unpopular. These candidates were the nominees of the two deeply unpopular political parties which have held a duopoly on the national political stage for years. There were 31.1 million votes cast in the republican primary and 30.6 million votes cast in the democratic primary. These each represent about 25% of the total votes cast in the national presidential election. With such a small number of voters deciding who will become the only two candidates with any real chance of winning, it is no surprise that so many voters choose to cast their vote for the ‘lesser of two evils’ or to skip the election entirely.

If you voted for Clinton, did you do so because you thought there is no other person country that you would have preferred for the job, or was your vote primarily out of opposition to Trump? The same question goes to Trump voters. If yours was an opposition vote rather than a vote that reflects your true preference, then you should seriously consider whether or not the electoral system makes sense. Even if you truly believe that there is no one in the country you would have preferred over President Trump, and you have no problem with the fact that he did not win a plurality of the popular vote, you should be aware that the shoe could very well be on the other foot in the future. 

This post discusses why a national popular vote is a preferable system to the electoral college, but I’ll emphasize that abolishing the electoral college is only a half measure. Plurality based voting systems (including a popular vote system) are among the worst of the dozens of systems ever proposed. Approval voting, range voting, and instant runoff voting are among vastly superior voting systems. These allow for a larger number of viable candidates since voters are allowed to provide information about multiple candidates instead of casting their single vote for just one. There should be a public conversation about which of these makes the most sense for the United States and I’ll hopefully get a chance to write about these systems some more in the future. However, a national popular vote would be an incremental step in the right direction and would take less effort to implement through the National Popular Vote Interstate Compact as opposed to an amendment to the constitution.

So let’s take a look at some of the arguments used in favor of the electoral college. As you read this, try not to think in terms of Clinton vs. Trump, but rather consider whether or not the electoral college honestly makes sense to you. There is no way to change the outcome the past election, but now is the time to think about how our voting system will shape future elections.

Note: the election data is taken from the county level vote tallies published by as of 21 Dec 2016. State level data is computed from the county level data for consistency with the county data, but these tallies may differ slightly from more recently published state level results. 

The Electoral College Ensures that Every State Gets a Say (it doesn’t)

A very common argument in favor of the electoral college is that it allegedly gives each state a fair say in the election. As the argument goes, the electoral college punishes candidates who appeal solely to populous states and ignore the needs of lower population states. This points to one of the main challenges in single winner elections: majority groups may consistently win over minority groups, effectively shutting them out of government. These minority groups may divide along lines based on geography, race, ideology, ethnicity, class, and so on. It is therefore desirable to have a voting system which requires winners to have broad appeal and penalizes a candidate for solely focusing on one group at the expense of another group. But does the electoral college actually accomplish this? First, let’s look at the number of total votes by state in the 2016 election.

Most votes are on the coasts, the Midwest, and Texas. There aren’t many voters around the Dakotas and Wyoming as well as parts of the Northeast and the South, but a majority of votes are not concentrated in any single geographic region. These votes are not counted nationally however, and each state casts its electoral votes for the popular vote winner in that particular state. The electoral votes granted to each state are partially in proportion to population. Each state is granted two electors regardless of population, and the remaining votes are allocated to states on a per population basis. Therefore one measure of the voting power of a particular state is the ratio of the electoral votes to actual votes of a state. If a candidate only needs to win over a small number of actual people to get a relatively large number of electoral votes, then smaller states become more valuable. Here is a plot of that ratio for the 2016 election.

Here we see that the relative value of many of the smaller states has increased. Artificially weighting votes by state is certainly one way to try to get smaller states some more power, but it turns out that the way electors are allocated is only secondary to the true problem of the electoral college. We can compute the popular vote outcome of the 2016 election if all votes, regardless of candidate, were weighted by state according to these proportions. This calculation just makes votes in small states worth more than votes in large states. In this case, the weighted popular vote would have still gone to Clinton, although Trump would pick up some ground. 

Republican Democrat Independent Total
Votes (millions) 61.14 62.31 1.37 124.83
Percentage 48.98% 49.92% 1.10% -
Weighted Votes (millions) 61.29 62.13 1.40 -
Percentage of weighted votes 49.10% 49.77% 1.13% -
Vote Difference (millions) 0.15 -0.18 0.03 -
Percentage Difference 0.12% -0.15% 0.03% -

The artificial weighting of votes is doesn't really change things a whole lot. However, winning a bare majority in the top 31 states ranked by their electoral to actual vote ratio would allow a candidate to win a majority 285 electoral votes with just 22.3% of the national popular vote using the 2016 data. How is that possible? The reason why is exactly the reason why the electoral college prevents all but a few states from having any meaningful influence in an election. To dig into this further, let’s look at how much campaign attention was paid to each state. collected the number of campaign events from the two main candidates by state. The distribution is plotted on the map below


Most states had no events, meaning that neither of the two major candidates felt they had to do any in person campaigning in most states. Many of those small states were completely ignored as well, with Vermont being the only small state that received much attention. On the other hand large proportions of events occurred in states like Florida, Pennsylvania, Ohio, and North Carolina. So what is so special about the few states that received nearly all the campaign attention? We can get some insight here by looking at the percent vote margin of the winning candidate for each state.

It should immediately become clear what is going on, the states that received the most campaign attention were simply the states where the election was close. Both candidates knew that it was not worth campaigning in any state that could not conceivably be flipped. This points to the primary flaw in the electoral college system. Electoral votes are awarded by a winner take all state level vote (with the exception of Maine and Nebraska), and if the state level race is not close, then that state is a waste of campaign resources. Therefore, the electoral college does anything but ensure that every state gets a say, in fact it ensures that only a very small minority of states are relevant in an election.

There are some additional consequences to this type of system as well. Since votes in only a few states are worth campaigning for, if a person moves to a new location, they may see their voting power increase or decrease as a result. For example, a resident of Clymer NY is in a state that is consistently won by Democrats by a significant margin. No candidate would ever have to care about this person’s preferences, since the electoral college ensures that no matter how they cast their vote, their state electoral vote will be given to the Democratic candidate. Drive a few minutes from Clymer to Wattsburg PA, and the votes of the residents there are now extremely valuable, as Pennsylvania was a hotly contested swing state which was crucial in the electoral college victory of President Trump. People may relocate for many reasons, including economic necessity or opportunity. Do you really think it makes sense to have a system that can cause someone to have such drastically different voting power if they relocate a few miles?

This demonstrates that the electoral college disproportionately empowers a small minority of states, but would a popular vote system do the same?

A Popular Vote Lets a Few Cities Decide Every Election (it doesn’t)

A common argument against a popular vote system is that the cities would decide every election, and rob large parts of the country of their voice.  To see if this holds up, let’s look at the county level vote data from the 2016 election. Below we have each of the 3200 or so counties in the US ranked by the total number of votes (note the logarithmic scale on the vertical axis). The green region represents a majority of all votes cast, while the green and blue region combined represent 80% of all votes case. 

This means that 50% of votes were cast in about 5% of counties and 80% of votes were cast in about 21% of counties. This is a pretty extreme distribution of votes, but this is fairly common for all sorts of different data. As an aside, the distribution of campaign events by state was even more extreme, with 80% of evens occurring in just 16% of states. This sort of 80-20 distribution was first observed by an Italian economist name Vilfredo Pareto, who showed that 80% of the land in Italy was owned by 20% of the population and that 80% of the peas he harvested from his garden came from 20% of pea-pods. This distribution is observed so frequently in population data that it may help guide our search for intelligent life outside out solar system. However, it also leads to a mathematical fact that reads something like a Lewis Carroll poem: The typical person lives in a populous county but the typical county is not very populous, therefore the typical person lives in an atypical county. And now we’re back to the problem of minority representation in a system where the majority rules. The preferences of those living in population centers could be very different from those in rural areas for a number of reasons, and generally city dwellers are more liberal than those living in less densely populated regions.

So what would it look like if a candidate won every single vote in those 5% or so of counties which hold a majority of votes with every other vote in the country going to the opposition? Here is the county map below. Even winning 100% support in coastal cities simply isn’t enough to win the popular vote, and there are a large number of states which have at least one county in the top 5%. Other than the fact that they are in the top 5% by vote count, many of these counties have hardly anything to do with one another. This reveals that the most extreme distribution of votes concentrated in the most populous counties in the country, a popular vote winner would not be able to solely focus on a limited geographical region, since the population centers in the US are dispersed across the country. Of course, no candidate can rely upon 100% support in any county, and in the past election president Trump won 23.9 million votes in these populous counties. Clinton did have 58.2% of the votes in these counties, which is a substantial win margin but it is no where near 100%. 

The fact that population centers are broadly distributed across the country and are not ideologically homogenous therefore forces a candidate seeking to win a popular vote election to have support that is broadly distributed across the country. This would spread campaign attention from the small handful of swing states that candidates targeted this past cycle, but it doesn’t solve the issue of representation of rural areas. While a popular vote winner cannot realistically win an election solely by winning majorities in population centers, a popular vote would allow scenarios where the winner loses in most counties. The alternative voting systems I mentioned earlier are part of the solution to this problem since they allow third party or alternative candidates to be competitive. I won’t go into too more detail on those systems here, but the Constitution offers another solution.

The federal government is composed of the executive, legislative and judicial branches which each have checks on the power of the others. The Senate is part of the legislative branch and consists of two senators per state, regardless of population. The combined population of the four smallest states of Wyoming, Vermont, Alaska, and North Dakota is about 2.7 million. This is about the same number of people that live in Brooklyn, yet there are two senators apiece these four states while the two New York senators must represent the interests of the nearly 20 million residents of the state. The number of seats in the house of representatives is more proportional to population, but not completely proportional. The executive branch can’t introduce legislation, and this gives smaller states a substantial amount of power.

This is a check on the power of the majority that actually works, but does the electoral college also check the power of cities? A good question is whether or not the electoral college would stop the election of our hypothetical candidate with 100% support in the top 5% of counties by vote count. Here is what that electoral map would look like. The blue candidate would win 308 electoral votes, enough for the presidency.

With the magic of the electoral college, the outcome of the election would have been exactly the same if this candidate had not won a single vote in any of the counties in states that went red. If these all of votes went to the red candidate in the first place, the blue candidate would win the electoral college with 2.5% of counties and 31.4 % of the popular vote. This electoral win comes from a subset (a little less than half) of counties needed for the most extreme vote distribution in a popular vote. This is why the notion that the electoral college protects rural voters from being trampled by city voters falls flat on its face. States with large cities can win a bare statewide majority for one candidate, and the votes of the residents of the rest of the state no longer matter. If enough states are won in this manner, a candidate can win an electoral victory with a massive popular vote loss by solely appealing to population centers.

This electoral map mainly consists of traditionally blue states along with swing states such as Florida, Michigan, and Pennsylvania. Texas is the only real red state on the map, and in the 2016 election Trump won it by over 800,000 votes. Texas turning blue seems unlikely at this time, but the point should be clear. The electoral college gives candidates a path to victory by appealing to population centers. 

This distribution of votes is illustrative, although not too realistic. With the actual data from the 2016 election, we can see whether or not Clinton’s electoral loss was the result of paying too much attention to cities. Some of the closest swing states were Florida, Michigan, and Pennsylvania, and these are states where a majority of votes statewide are from counties in the top 5%. Florida plus at least one of the other two would be sufficient to flip the electoral college. The republican margin in the second column is the statewide margin. The remaining columns are only for the votes cast in counties which are in the top 5% by number of votes nationwide. 

State Rep. Statewide Margin Rep. Votes* Dem. Votes* Ind. Votes* Margin as % of all votes*
MI 11612 1406392 1025092 7845 0.48
FL 119770 3377838 2820821 17360 1.93
PA 68236 1866840 1290228 9928 2.15
WI 22871 585692 339882 3852 2.46
* these totals are only for counties which are in the top 5% of counties which account for a nationwide majority of votes

There are more than enough votes in just these counties for the democrats to tip the scales state wide by increasing turnout and flipping republican and independent voters in just the largest counties in each state. The electoral college allows the Democrats to pursue a strategy where they aggressively pander to voters in a small fraction counties (which happen to be among the most populous in the country) in a small fraction of states. Pandering to this small minority of voters could cost them millions of votes in other states and even cost them the popular vote, and it wouldn’t make a difference as long as they keep enough votes to win bare majority in enough states for the electoral win. Both parties are surely aware of this information, and they will campaign accordingly in the next election. This all but guarantees that once again only a small number of states will receive any significant campaign attention.

The United States is a Constitutional Republic, not a Democracy (that’s true actually)

Democracies and Republics are both forms of representative government, but there are some differences between the two. The difference as it pertains to the US is mainly that states are represented in institutions of the federal government, in addition to governmental bodies that represent people. The US constitution defines the powers and responsibilities of state and federal governments and the relationship between these entities. Article II of the constitution defines the process of electing the president by the electoral college. This grants the states a troubling kind of power that no one should be comfortable with. States are permitted to cast their electoral votes essentially however they see fit, and most states do this in a winner take all system. Since electoral votes are not allocated in proportion to the state popular votes the constitution grants the state the power to disregard your vote, and give it to someone else. The state level winner does not even need a majority of votes, and can win 100% of the electoral votes for that state with just a plurality.

As we saw, only a few states were close races, but even states that are dominated by one party or another are often extremely heterogeneous. We can visualize this by plotting the “lost counties” as I’ve decided to call them, or counties where the county winner was different than the statewide winner. Red counties are counties that went for Trump in Clinton states, and blue counties are counties that went for Clinton in Trump states. 

There are hundreds of lost counties over the country for both Democrat and Republican voters. Many lost counties even form contiguous regions or clusters. Republican lost counties are generally rural areas in states with large cities, Democrat lost counties are generally population centers and areas with large minority populations in more conservative states. The borders of states are arbitrary in the sense that to do not form coherent voting blocks where the residents have similar preferences. The same argument applies to counties as well, and even though a county may have voted with it’s state, the losing candidate still may have had some significant support. Below is the distribution of total lost votes by state.

Millions of votes were lost in red states and blue states alike. The national totals of lost votes are below

Republican Democrat Independent Total
Lost Votes (millions) 20.97 31.71 1.37 54.06
Percentage 34.30% 50.90% 100.00% 41.53%

The electoral college allowed states to change the votes of over 54 million voters. A majority of Clinton’s votes actually came from states that went to Trump and were therefore lost. This makes it clear why it is possible to win the electoral college while losing the popular vote. For example, the 3.8 million votes were cast for Clinton in Texas, yet not one electoral vote went to Clinton in Texas.

If you are pleased with the outcome of this election because you think that Donald Trump is the best possible person in the country for the job of commander in chief, then you should realize that our constitutional republic grants the government the power to take away your vote as well. Maybe you were among the 21 million Trump voters who had their vote changed to Clinton, but it worked out for you in the end anyhow. Keep in mind that there are scenarios where Democrats could win the electoral college with the support of cities and a minority of the popular vote, reversing the roles in this election. When the government has the power to change someone else’s vote, they have the power to change your vote as well. Is this really the type of power you trust the government with?

The constitution has been amended many times to restrict the power of the government and expand the right of the people, and an amendment is what would be required to abolish the electoral college. Fortunately, a popular vote can be implemented without the need for an amendment. The National Popular Vote Interstate Compact is an agreement among several states to cast their electoral votes for the winner of the national popular vote, regardless of the state level vote. For a state to adopt the compact, must be passed by the state legislature and signed into law by the governor. Check here to see the status of your state. If you live in a state that is not part of the compact, writing to your state representatives is a simple action you can take immediately. If you have never written your representative before, or you do not know who your representatives are, a site called has simplified the whole process. Here is what to do:

1. Go to and enter your address. This will take you a page with a drop down menu listing your state and federal representatives.

2. Under pick an issue, choose "write your own". Writetocongress has form letters for a number of issues, but at the time of posting, they did not have anything for this topic.

3. Choose your state official from the second drop down. Since you have multiple state officials you can write multiple letters, and writetocongress will take care of generating the letter head for you.

4. Click the letter template to edit the body of the letter, and urge your representative to support the national popular vote interstate compact. I've taken a crack at writing a template which you can use or modify as you see fit.

5. Print the letter and send it via snail mail.

And there you have it. You can let the people with the power to legally join your state with the compact know your opinion, and it wont take much more time than posting a tweet.

If you live in a state that is already part of the compact, you shouldn’t be satisfied with that. The compact takes nullifies the ill effects of the electoral college, but it still leaves states with the power to change your vote. Research alternative voting systems such as approval voting, range voting, and instant runoff voting. I guarantee that regardless of your politics, you will find these systems a great deal more sensible than the electoral college or a plurality based national popular vote.

Lastly, if you are a still a proponent of the electoral college, I am open to being convinced. Tell me why it makes sense for votes to be weighted differently based on voter’s geographical location, why it makes sense to aggregate votes by state which are often highly heterogeneous regions, why it makes sense to allow candidates to compete for votes only in a handful of swing states, and why you believe the government should have the power to change your vote if it is different than a plurality of voters in your state.

Interested in how the data in this post was gathered, analyzed, and plotted? You can find the full source code on GitHub. The source code for the script for collecting the county level vote data from can be found in this project by Tony McGovern.