By Mike Koetting July 11, 2019
In almost everyone’s list of the reasons that Hilary Clinton is not the president of the U.S. is the Electoral College. As we all know, Clinton won the popular vote by almost three million votes (2%) but lost in the Electoral College, 304-227. The usual argument suggests the number of low-population red states–that get disproportionate influence because of the way votes in the Electoral College are distributed–threw the election to Trump.
This argument, however, is wrong.
To be sure, the structure of the Electoral College is goofy. There are wild disparities among states as to how many voters are represented by each electoral vote. It ranges from a low of 193K in Wyoming, to a high of 755K in Texas. In other words, each individual voter in Wyoming has almost four times the influence on a single vote in the Electoral College as a vote in Texas.
But it turns out this didn’t make much difference in the 2016 Presidential Election. Specifically, suppose you assume the 538 electoral votes were distributed not as they are currently (number of Representatives plus number of Senators) but were distributed proportional to the actual population of each state. I did this exercise using 2018 population, but the results would not have been appreciably different using 2016 data. In my case, Trump still beat Clinton in the Electoral College by almost the same margin, 305-233.
I think a more accurate assignment of the problem is the “winner take all” (WTA) approach that is used in 48 states. Under this system, the candidate getting the largest number of votes gets all the electoral votes from that state. This approach is not mandated by the Constitution and is solely at the discretion of the states. WTA rules have been common since the mid-1800s, and the appeal to political parties is obvious—it maximizes their clout.
I conducted a second counter-factual exercise on the 2016 voting data, this time assuming that in each state the electoral votes (as they currently stand based on 2010 census) were allocated not WTA but proportionate to the actual votes for the candidates in the state. In this case. Clinton had a slight margin, 269 to 266.
This would not, however, have made Clinton the winner because the presidency requires a minimum of 270 votes. In this case, the election would have been thrown into the House of Representatives, where each state would get one vote. This in turn would most certainly, once again, result in a Trump presidency.
To be fair, in a margin as close as my proportional vote approach, the final result depends to some extent on the rules used for allocation, including third party candidates. While I would argue the rules I used are reasonable (see note at end concerning data), I would not be at all surprised if real world, on-the-ground applications of this general principle would change the totals by enough to change this outcome. But since this exercise is entirely hypothetical, the exact results are not as important as the general principle: a WTA approach can distort the connection between popular votes and electoral votes—to no particular policy end.
A desire to improve the connection between the popular vote and electoral votes has led to the National Popular Vote Interstate Compact. This is an agreement among a number of states that as soon as enough states have signed on to the agreement that a majority of electoral votes (270) would be impacted, the states would agree to award all their electoral votes to the winner of the national popular vote. This approach does not require a constitutional amendment and would, as long as it was in effect, hardwire the connection between national popular vote and the presidency.
It seems virtually impossible the Compact could be in effect for 2020. At present, states with a combined total of 196 electoral votes have signed on, primarily on the east or west coast. There are a few states where one can imagine adoption in the near future, but those yield less than half of the 74 additional electoral votes needed to get to 270. Conversely, it is hard to imagine most of the other states signing on because they voted for Trump in 2016 and recognize that Trump’s most likely road to re-election is through the Electoral College rather than the popular vote. While there is reasonable evidence that over time this approach would not favor one party over the other, long-run views are hard to come by in this day and age.
I support this approach, although it is not without concerns. For one, it absolutely removes third-parties from the board. In 2016, almost 6% of votes were for third-party candidates. I suspect different people have different attitudes on this, but there can be no question that third parties have had significant impacts on elections, and arguably a growing impact. Ross Perot was instrumental in electing Clinton over Bush One and the Ralph Nader campaign almost certainly cost Al Gore the presidency. In 2016, Johnson beat Stein by a 3-1 margin in Michigan, Pennsylvania and Wisconsin, but in all three states if Stein voters had instead voted for Clinton, she would be president of the U.S.
The bigger issue for the Compact to overcome is psychological: are the citizens of the non-metropolitan areas ready to come to grips with the fact that residents of the big-cities outnumber them significantly, and, accordingly, might out-vote them?
The National Popular Vote campaign argues that their approach, since all votes are equally important in the final total, will serve as a force to equalize campaign focus. This argument seems to me to be far-fetched for logistical reasons and not that important for conceptual reasons.
Logistically, when the only thing that counts is votes, campaigns will as much—perhaps even more—focus on high population areas. In Harris County, TX (Houston), the number of potential voters who didn’t vote is equal to the entire populations of Wyoming, Alaska and South Dakota combined. Both Clinton and Trump got well over 500,000 votes in Houston, so cutting even a bit into the other’s margin is a rich vote opportunity. I believe candidates will do the same math. But this get back to the question that is at the heart of the theoretical debate. What do we think democracy means? Under what circumstances is a democracy better served by having a president elected by a minority of the voters?
There is no question the Founding Fathers made a deliberate decision to weight votes by states. But I don’t think that is needs continue to be the case. That approach was a compromise based on necessity. At the time, there was no America; only a collection (at best a confederation) of individual states. There would have been no America without that compromise. But in the 250 years since then, the needs for this compromise have dimmed and the distortions caused by it have increased as population in big cities has increased faster than the rest of the country. Attachment to states have waned as mobility and communication have made the culture more national than regional. Campaigns are increasingly national in orientation. There may be differences in ground game and get-out-the-vote organization, but those are more a function of national fundraising than anything else.
There are no longer compelling reasons to privilege differences among states as more important than any of the other differences that might arise. For instance, the number of non-Hispanic blacks in Cook County alone is greater than the population of 9 states and the District of Columbia. While their interests are clearly different from the white or Hispanic populations, no one is suggesting they get their own electoral votes. Making all votes equal is probably the best we can do. And, actually, is a pretty good goal.
NOTES ON DATA
I used the vote totals as reported on Wikipedia.
In redistributing electoral votes by actual population, I used 2018 population to spread 538 votes proportionately. Note the hypothetical result I have shown totals to 538 votes, the total allowable. In fact, in 2016, there were only 531 electoral votes cast. (I have no idea why those electoral voters didn’t vote.)
For the proportional distribution within state, I used the current electoral votes by state (i.e. those set based on 2010 census). I rounded totals to allocate all the electoral votes. In Maine and Nebraska, the only two states not using a WTA approach, I left the results alone. I ignored third party candidates unless they received a share of votes equal to the proportionate share of one or more electoral vote. With this rule, the scenario awarded three electoral votes to third party candidates, resulting in a total for Trump and Clinton of only 535 votes.