The 2016 Presidential election may be the most discussed election in American history. From Democrats almost neurotic analysis of how Hillary Clinton—one of the most qualified Presidential Candidates in history—could have possibly lost to Donald Trump—a political new comer; to Russian interference, election security, and the Muller Investigation; to Trump’s insistence that he won the popular vote but for the “millions of people who voted illegally”, without any evidence to support this claim. However, Trump’s defensiveness over his popular vote deficit is revealing of a feature within our electoral system. Namely, that the Electoral College (EC) over weights the votes of less populated states making it possible for a candidate to lose the popular vote by 2.8 million votes but still win the EC, as Trump did in 2016.
To be clear, this is a feature, not a bug. When shaping our government the Founders purposefully created a system which would weigh towards states with lower populations. Fearing a tyranny of the majority, the Framers guaranteed all states a minimum of 3 electoral votes, in order to protect the interest of those less populated states from being dominated by more populated states. In today’s context this means that Wyoming, a state with the lowest population of voting aged adults in 2019 according to Census data, accounts for 0.17% of the total voting aged population in the US, but 0.56% of the total EC votes. Conversely, California, with the largest population of voting aged adults, accounts for 12% of the total voting aged adults in 2019 but only 10% of the EC votes. With a large and diverse country, this may be a good feature of an electoral system; however, the disconnect between the popular vote and the EC seems to have flipped the Founder’s concern of a tyrannous majority on its head, with an entrenched minority able to punch above its weight to dictate national politics.
My intent is not to call into question the legitimacy of Trump’s election. Both Trump and Clinton were playing by the same set of rules and Trump won fairly under those rules. However, I think it is worth revisiting those rules which govern our elections, specifically the pluralistic system of allocating EC votes, by looking at the results and outcomes of our elections. Using election results from the Federal Elections Commission, in 2016 Hillary Clinton won 48% of the popular vote versus Donald Trump’s 46%, but due to our ‘winner take all’ electoral system, that translated into Trump winning 59% of states and 57% of the EC compared to Clinton’s 41% and 43%, respectively (see Figure D, below). This is not the only time in our history that we have seen a disconnect between the popular will of the people and the results of the EC. As shown in Figure A, below, in the five elections which took place this century, two have resulted in victors with a negative margin of victory; 20% of our elections this century have been decided by a minority of the population.
Figure A:
Intuitively, I think one would look at the results of the 2000 and 2016 elections and say the outcomes are undemocratic based on the results the popular vote. Conversely, one might understand the Framer’s intent when adopting the EC System, as outlined above, and would accurately point out that the system is working as intended. However, those same Framers warned against political parties, and certainly did not foresee power concentrating within two parties as it has today. Referring to Figure B, below, we see that even in years that the governing party does win a majority of the popular vote, it is only by a slight majority—3% at most.
Figure B:
In other systems of democracy this would push the ruling party toward building a coalition to effectively govern, but in a highly partisan, bipolar government that simply is not an option. One might not be all that surprised by the results in Figure B given the deep partisan divides in the country. Though, what is more striking about the data above is not the results for the two parties, but the ‘None’ column. In 2016, enough voters—28,863—voted for “None of these candidates” for this to register, ever so slightly, with 0.02% of the total popular vote. What’s more surprising, this voting option registered despite Nevada being the only state having this ballot option. In that state, “None of these candidates” garnered 3.9% of the popular vote. This is a fairly damning indictment of the state of the nation’s two parties from the voters of Nevada; it is an indictment from the voters who have the option to vote their displeasure with the current status quo. If we were to rerun the election results to include reported registered voters from the Census Bureau the picture looks much worse.
Figure C:
The data in Figure C, above, was calculated using election results from the Federal Elections Commission and reported registered voters from the Census Bureau. Using this data we can see that on average over the last 5 elections 16% of registered voters did not cast their vote. This should register as a vote for “None of these candidates”, though this context is almost never provided when covering election results. In fact, it is not included in the topline numbers when looking at election results on the FEC website. This picture of American political participation would only get more grim if we were to run the same calculation for all voting aged citizens. It is worth also considering that these calculations were run in Presidential election years and for Presidential elections, which tend to have the highest voter turnout.
Political participation is an important issue to consider and efforts to enfranchise more voters should be taken seriously. Nevada’s concept of providing an option to unequivocally state dissatisfaction with the two parties rather than voters staying at home and being branded apathetic seems a simple measure to take. Making election day a national holiday so people do not have to take PTO to vote seems another intuitive fix. These, however, are by no means comprehensive and much more detailed policy proposals exist to engage those citizens who have historically not participated. These policies should be championed.
So, what we have seen so far is that our current electoral system is (1) not responsive to the plurality of the electorate in all cases and, (2) is not engaging citizens to be politically active.
First, to address the responsiveness of our system to the popular electorate, Figures B & C, above, show the election results as total votes and percentages. Figure D bring these together to highlight the disconnect.
Figure D:
In the 2000 election, the difference in the popular vote was 543,895 votes, or 0.5% of total votes cast, so when looking at the electoral results there does not appear to be a deep disconnect. But in 2016, when that deficit grew to 2,868,686, 2% of total votes, the issue became more clear. Looking into the results by state in 2000 six states were decided by 1% or less; Florida (25), Iowa (7), New Hampshire (4), New Mexico (5), Oregon (7), and Wisconsin (11). In all of these states, the winner also did not win a majority of votes cast, and that is 59 EC votes allocated entirely to a candidate who did not win a majority in these states. Similarly, in 2016 six states were decided by 1% or less; Florida (29), Michigan (16), Minnesota (10), New Hampshire (4), Pennsylvania (20), and Wisconsin (10). Again, in none of these cases did the winning candidate garner more than a plurality of votes cast. These are not inconsequential, small EC states in all cases. It should also be understood that this cuts both ways; Gore in 2000, and Clinton in 2016 both hauled in EC votes this way despite losing the EC. Appendices A & B should be referenced for the full results of the 2000 and 2016 elections.
To take a step back and not focus for a minute on the specifics of these two elections, let us instead look a hypothetical scenario. One could imagine a case in which a candidate wins a majority of the states, let’s say 32+, and so wins the EC in a landslide. In this scenario this candidate also wins the popular vote. The House and Senate are swept as well. One party controls the government. The candidate would inevitably enter office declaring a mandate from the people to fully enact their policies. But, in this scenario, this candidate only receives 48% of the popular vote; a plurality, not a majority. Why then does our electoral system recognize this as a mandate? A majority of Americans have not expressed their support for these policies and yet our electoral system warps and amplifies election results, multiplying, and in some cases flipping, popular vote margins.
A way in which to mitigate this distortion, while maintaining the EC would be for each state to apportion their EC votes based on the percentage of votes cast for a candidate. Admittedly, when working with specific elections, one can shift the allocation to achieve a specific, desired outcome. The allocation ranges I applied are as follows:
We can argue with the specific allocation percentages, but when allocating EC votes we can pull the EC margin in line with the popular vote margin, as show in Figure E, below. Further documentation of the allocation is included within Appendices A & B as well.
Figure E:
By aligning EC results with the popular votes in this way we have also made our elections more responsive to voters, which means that individuals’ votes matter. Under this system Trump picks up 19 votes from Republicans in California and Clinton pulls in 17 Democrat votes from Texas. All of a sudden those voices in those states are heard. Therefore, a Republican voter in California may understand that casting their vote will not result in an outright victory for Republicans in a statewide election, but their vote may result their chosen candidate receiving a larger percentage of EC votes. If voters believe their vote will have more of an impact, then they will be more likely to turn out and support their candidate. By addressing the responsiveness of our electoral system to the proportion of votes received we are also addressing voter engagement.
It is also worth noting that when the US Constitution was originally adopted in 1789, Article II, Section 1, Clause 3 described the election process for the President and Vice President. Much like today, the candidate that receives the most electoral votes was elected President. However, the President and Vice President did not run on a single ticket, the Vice President was elected as the runner up candidate. This was amended 15 years later when the 12th Amendment was adopted, changing our elections to elect both the President and Vice President on the same ticket as we do today. This is of note for a couple reasons. First, this is an example of proportionalism in the history of American elections. I am by no means suggesting we should revert to this specific system of a proportional elections; it could lead to the Presidential and Vice-Presidential offices held by opposing parties, and likely result in severe disfunction. It is for this reason that the 12th Amendment was originally adopted. Second, there has been a willingness, in our history to alter our election processes. We should have the same willingness today to adopt solutions if we identify issues.
In addition to examples of historical alternatives, we have contemporary alternatives we could examine as well; namely the Congressional district method of allocating EC votes used Maine and Nebraska. Both Maine and Nebraska award 2 EC votes to the winner of the statewide popular vote. The remaining EC votes are awarded by district, with the winner of the plurality of the popular vote in a district winning that district’s EC vote. This could get us closer to solving the issues identified above, however, it would also create greater incentives for gerrymandering of districts to benefit a single party as they would reap the rewards in both Presidential elections and House of Representative elections. This brings us to our next subject.
To this point we have only examined Presidential election results, and there is reason to believe this change would only do so much to activate voters currently sitting on the sidelines. Based on past Presidential election results, supporters of third-party candidates may still see votes for their preferred candidate as throw-away or protest vote. Despite the increased odds that third-party candidates will win EC votes under the proportional system, it would still be very difficult for them to reach the 270 EC summit to win the election. However, there is an additional avenue we can engage with these voters. If we look at US House of Representative election results we see an issue similar to the EC results. In 2016 Republicans won 48% of the popular vote in House elections but 55% of seats. That same year, Democrats won 47% of the popular vote, but only 45% of seats. In 2018 this benefit flipped; Democrats won 53% of the popular vote and 54% of seats, Republicans won 44% of the popular vote and 46% of seats. It should be pointed out that third-party candidates won 4% of the popular vote in 2016 and 2% in 2018 though they won no seats in either election. In 2019 Representative Justin Amash left the Republican party to join the Libertarian party and is the lone Independent member of the House, but he was elected in 2018 as a member of the Republican party. To be clear, at the state level, both Republicans and Democrats benefit from districting which allows them to pick up more seats than the popular vote would suggest. See Appendices C.1 & D.1 for a state by state break down of these margins—it should be noted that these appendices only include states with more than one Representative. Gerrymandering is an obvious blight on our democratic system as its intent is to dilute certain votes to the point of irrelevance. One solution is to abolish Congressional districts.
Instead of a state being broken into a multiplicity of Congressional districts with one Representative for each district, voters would cast their ballot for a party and seats would be allocated based on the percentage of the popular vote won. Without Congressional districts, parties would be unable to divide voting blocs through gerrymandering in order to dilute those voters’ power. Admittedly, this would decrease regional representation, but it would increase political or ideological representation as voters would not be forced into a decision between two parties.
It is important to consider how much regional representation the current system provides though. Let us quickly take a look at the 116th Congressional district map for my home state, Ohio. After the 2010 census, the Ohio State Legislature, held by Republicans, redrew the Congressional districts for the state which resulted in 9th District absorbing part of the 10th District of the 111th Congress. This caused legacy 9th District Representative, Democrat Marcy Kaptur, to run against legacy 10th District Representative, Democrat Dennis Kucinich, in the 2012 election for the 112th Session of the House of Representatives. Dennis Kucinich, who was a popular Democratic Representative, was redistricted out of the House of Representatives. Today, Ohio’s 9th District stretches from Toledo in the west to the westside of Cleveland in the east—two of Ohio major metropolitan areas, separated by over 100 miles and almost a two-hour drive. This district spans five counties. Technically contiguous, but entirely convoluted. Ohio’s 11th District does not quite pass the eye test either, extending from Cleveland down to Akron. These two cities, however, are at least in the same Combined Statistical Area as designated by the Census Bureau and separated by fewer than 50 miles and less than a 45-minute drive.
In the case of the 9th District, are the constituents truly afforded regional representation? Sure, all have interests in policy which protects the Great Lakes as historically high-water levels in recent years have resulted in erosion of the shoreline and annual deadly algal blooms result in boil-water advisories for coastal communities. However, this district spans from rural agricultural communities, to exurban, to suburban, to urban communities—many of which are still heavily manufacturing based. The interest of each of these groups is going to vary depending on their geographic and demographic needs. Any congressional district is going to include some mix of these geographic and demographic groupings, but if one of the benefits of Congressional Districts is regional representation, then it would seem that this benefit is being nullified by the current system of gerrymandering.
Returning again to the House elections more broadly, in Appendices C.2 & D.2 we can see what the House election results may have looked like in 2016 and 2018 had we apportioned seats based on the popular vote. In 2016, the percentages called out above translated into Republicans winning 28 seats more than they would have if apportioned; Democrats won 17 fewer and Independents won 11 fewer. In 2018, Democrats won 9 more, Republicans 1 fewer, and Independents 8 fewer.
One could imagine that after a few cycles under this system the number of third-party Representatives might increase. Members currently identifying with one of the two dominant parties might leave as they would longer be constrained by the confines of the existing party structure. Ultimately it would recognize a reality that currently exists within Democratic and Republican parties; these are already coalitions of diverse ideological groupings, but given the power dynamics they do not necessarily need to be as responsive to all members of the coalition. Or, more insidiously, these parties are captured by highly vocal and politically active members of the party who tend to be more partisan and extreme. The centralized nature of the two parties also allows large money donors, PACs, and “dark money” contributions to focus their efforts to the greatest effect. If you break up these parties, the money is diluted, the power weakened. Similar to the discussion of voter engagement I am under no illusions that this is nearly sufficient to address the issues of campaign finance, but it may be a back-door start.
I had mentioned earlier that it is relatively strange that the winner of our elections come to power without a majority of support, but govern as though they had. No work is done to build a coalition to represent a majority of Americans because no work needs to be done. Under the system proposed above, even if there is not substantial movement in the proportion of votes cast for the two major parties in Presidential contests, if the House of Representatives is disaggregated among multiple parties then coalition building will be necessary to govern. Consequently, policies will be amended to reflect that popular majority created. Cross party cooperation will be a necessity to effectively govern, and Representatives will be able to more freely switch between traditional Republican and Democratic groupings on a policy-by-policy basis. In short, the hope is that there would be increased cooperation within the House of Representatives.
At this point one may be wondering what my proposed solution for the Senate is, and unsatisfactorily perhaps, I do not have one here. Similar to the concern described above when discussing the weight of EC votes, the senate is much more weighted toward lower population states as every state has an equal 2% representation, despite 0.17% - 12% of voting aged population gap between Wyoming and California. However, this has historically been a workable system and disenfranchising low population states is a concern. The hope would be that by introducing proportionality to the EC in Presidential elections and the House of Representative elections, third-parties would be legitimized and the two parties’ hold on the Senate would be similarly weakened. Interestingly we see more Independent affiliated Congress-people in the Senate than in the House of Representatives; Bernie Sanders from Vermont and Angus King from Maine. It is possible that with a more diverse Senate, coalitions of greater than 60 members could be created to overcome the filibuster, enabling effective legislating and governance. However, it is equally likely that this 60 member threshold would be forever out of reach and the Senate would be in perpetual deadlock. This may be romanticizing the legislative process, but I think it is important for a legislative body to be afforded the time to deliberate legislation, and fully consider the effects; reading ‘Green Eggs and Ham’ on the Senate floor is not that. At the very least, some reform of the filibuster is needed, but I am not sure outright abolition is the solution.
With redistricting on the horizon as we complete the 2020 Census, now seems to be an ideal time to adopt the proportional system laid out above to avoid potential for gerrymandering by either party. As stated at several points above, this is not a holistic remedy for the issues ailing American democracy, we still need specific approaches to address voter enfranchisement and participation, campaign finance, and Senatorial procedures. And those are just the issues mentioned herein. I do, however, believe that this would help to address voter participation and would increase the political and ideological perspectives represented within our government, and would give us a better chance to identify those issues and work to amend them.
See the Electoral College Proportional Analysis spreadsheet for Appendices A & B
See the House of Representatives Proportional Analysis spreadsheet for Appendices C & D
Figure D – Ohio 116th Congressional District Map from Census.gov
