Garett Jones  

Why I Love the Electoral College

Two Verdicts on Two Replies to... Bruce Bartlett's Mistake...
There's some evidence that democracy itself makes people happier, but largely I see democracy as a means to an end.  One among those ends is "reducing social conflict."  

The electoral college, set forth in the U.S. Constitution, is a great tool for reducing social conflict across regions of the United States.  You might think that's a crazy claim--don't we see maps of red and blue, and aren't the red places--the places supporting the Republican--mostly in the South and Midwest?  Indeed, and that pattern across regions is key to explaining how the electoral college defuses some social tension.  

As it stands now, the states implicitly vote for the President.  Each state is granted a number of electors (equal to the number of House members plus Senators), so populous states states get more weight.  In almost every state, every single elector votes for the candidate who wins the plurality of that state's popular vote.  

That means candidates only care about winning a plurality of the votes in each state---winning California by one vote is just as good as winning by two million.  Of course, there's always some uncertainty about how things will turn out, so candidates love a cushion, but it's safe to say that if your state is polling 65% for a particular presidential candidate, neither candidate is likely to campaign there any time soon. 

And that's great news for social peace.  We rarely hear too much about regional issues in the U.S. other than farmers vs. everyone else.  But if the presidency was decided by majority rule, I'm sure we'd hear a lot more about regional differences.  Could a presidential candidate get 75% of the votes in Texas, Louisiana, Mississippi, and Florida by promising broad-based Gulf Coast subsidies and a few other goodies?  Could a candidate get 85% of California's and New York's votes partly by offering housing subsidies for people facing high housing costs?  

I don't know: But if we got rid of the electoral college and had a popularly elected president we'd sure have a chance to find out. 

As it stands, presidential candidates are trying to appeal to the median voter in each state across a large number of states.  That's how you get to be president.  This reduces regional tensions because candidates are never trying to get 90% of the votes in a state.   When you're pitting 90% of one region of the country against 90% of another region of the country, you're substantially raising the probability of social conflict.  

Too many civil wars are based on regional differences for this to be no big deal.  And you don't need to get to the point of civil war to get bad outcomes--mere regional transfer programs, switching across regions every four or eight years, would be quite bad enough. 

Right now, U.S. presidential candidates have zero interest in winning 100% of a state's votes.  But I'm guessing the campaign consultants could find some underexplored regional tensions if the incentives were right.  

Here's hoping they don't get that opportunity

Comments and Sharing

CATEGORIES: Public Choice Theory

COMMENTS (30 to date)
gwern writes:

> As it stands, presidential candidates are trying to appeal to the median voter in each state across a large number of states. That's how you get to be president. This reduces regional tensions because candidates are never trying to get 90% of the votes in a state. When you're pitting 90% of one region of the country against 90% of another region of the country, you're substantially raising the probability of social conflict.

Isn't it more like, they are trying to appeal to the median voter in a few specific swing states which will included and get their overrepresentation by sheer geographic size, thereby essentially removing entirely the preferences of large chunks of the population? (Enjoying your ethanol subsidies?)

And I suppose appointing a single voter as dictator could reduce tensions even further...

Daniel Artz writes:

Gwern, it is of course true that today, candidates are focusing only on a few key swing states. But ask yourself wether or not that would continue to be true if either candidate, in an attempt to garner some of those key swing states, took a position that was likely to aggravate his or her "base" in a safe state. While the "safe" states, like New York and California for Democrats,and Texas, etc. for Republicans, might not be truly "in play" for most elections, the concerns of those states, and issues capable of moving those states from "safe" to "in play", serve as a limit on just how far either candidate can go in pursuing the swing states. Remember that "safe" states can and have switched - before 1960, Texas was almost always "safe" for the Democrats, and Democratic-controlled legislatures in Texas continued to keep Democrats in control of the Texas House Delegation into the 1980s through dome very creative gerrymandering. Tom Delay's work on redistricting in Texas after the 2000 Census did not take place in an historical vacuum - it was payback (over-agressive, perhaps, but payback nonetheless) for decades of Democratic gerrymandering which sought to minimize the election prospects of Republicans. Just because a state is "safely" in the column of one party or the other does not render it irrelevant - those safe states create meaningful limits on the ability to pander to the swing states.

Daniel Kuehn writes:

gwern - I don't think that's the right way to look at it. Just because some states aren't contested doesn't mean that the candidate isn't appealing to the median voter in the state. Without the Electoral College they'd be seeking out the median voter in many fewer states.

Garrett - I strongly agree. This was my homage to the Electoral College a couple weeks ago:

ssh writes:

I am just glad that I don't have to see any campaign commercials, given that I am in Texas. That alone justifies the electoral college in my book!

(Of course, immigration is going to change that, probably in the next decade or so)

Alan Shields writes:

One point Dr. Jones does not bring up is the EC's role in reducing incentive and opportunity for electoral fraud in the Presidential election: it's easiest to commit voter fraud where your candidate has high support already - what's a few more people voting how everyone expected? EC makes these extra votes irrelevant. Where it's most worthwhile to perpetrate fraud is in close contests, and in those the state is likely to have divided control and close monitoring.

This doesn't mean it doesn't happen - the business in 2000 in Ohio was certainly suspect - it just means it's harder to do anywhere that it might matter.

Andy Hallman writes:

Hi Garrett. This is some interesting speculation, but really nothing more than that. I'd like to see if regional differences are larger, and cause the kind of strife you're talking about, in countries that elect their presidents by popular vote, which is most democracies.

The only difference I can see between a popular vote referendum and the electoral college is that the candidates would spend more time and energy in big cities across the nation rather than in canvassing small towns in the swing states.

Regionalism matters less than you think it does. Some time ago on this blog, Bryan posted poll results indicating that even people in New York City favored farm subsidies to make food cheaper.

Nathan writes:

I'm not sure I buy the argument. To me, the EC seems to create a situation where candidates pander to small constituencies in swing states to the detriment of everyone else. Consider our ridiculous 50 year embargo on Cuba. Supported by a small number of anti-Castro diehard Cubans in Florida, generally opposed by everyone else. But because the Cuban vote is enough to swing a closely divided and hugely important state like Florida, the ban continues. The same can be said for tariffs that protect certain key industries, like cars in Michigan or tires in Ohio. Everyone else gets screwed so that the candidates can scoop up a few percentage points in key states.

Henry writes:

In the only empirical test of your hypothesis (the 1860 election and the Civil War), the Electoral College performed in the exact opposite way, as Lincoln won by moderate margins in the North while receiving almost zero support in the South. Though he did win a plurality, he would have won even if all other candidates combined their support.

This highlights the flip-side of the argument - the Electoral College gives you no incentive to try to win votes in states you have no chance of winning. Frankly, I find that prospect far more worrying for regional strife than your concerns.

As for the pondering of the counterfactual, we have some evidence for what would happen - namely, every single other election ever. Do gubernatorial candidates, for instance, direct all their focus to pandering to their strongest geographic bases? Not that I can see. There might be geographic splits (see Illinois for a notable example), but no-one wins by completely writing off the regions they are behind in.

Bob McGrew writes:

Living in California, I also thought about the downside that the electoral college over-represents the interests of swing states.

The advantage Garrett cites is actually just the advantage of voting by district. If states apportioned their electoral college votes by congressional district (as I believe Nebraska already does?), we wouldn't have the downside of running up the vote regionally, but there would be enough competitive districts that every state would get attention.

Of course, this also requires a limit to gerrymandering, so this is all in the world of ideal politics anyway... but it's useful to understand what could be better and how. Every once in a while, you get the chance to do it.

Tom E. Snyder writes:

The electoral college of today in no way resembles the electoral system of the US Constitution. Electors were not to be sent to vote for a predetermined candidate but to chose a president and vice president. Today's electoral college could be improved but I don't know that it would make a lot of difference since the 2 main choices are really selected by an oligarchy.

Tom West writes:

Not much to add, but I hadn't thought of benefits of the EC in the terms that Garrett mentioned. I found the post very interesting.

(As an aside, watching American politics did teach me that having politicians actually be responsive to their constituents (as opposed to party loyalty trumping all), while great in theory, has some rather large downsides in practice.)

jstaples writes:

The electoral college makes perfect sense when viewed in the context of a constitutionally restrained federal government.

A strong state and restricted federal government allows the population a more powerful voice in matters that directly concern them. It is much easier to get the attention of state politicians than the federal bureaucracy.

In that type of system the president would have far less control and influence over the lives of ordinary citizens. His proper constitutional authority would be much more limited and affect the state government very directly while affecting the average citizen only indirectly.

In that context, it is entirely logical for each state to vote as a bloc on which candidate best represents the state interest. The distribution of votes in the electoral college was meant to provide a method for allowing more populous states an increased representation without completely minimizing the influence of smaller states.

In our current state of affairs the presidential election directly affects minute aspects of our lives --- for some people right down to how much money the government gives them to sustain themselves. Because of that I'm not entirely convinced the electoral college is still serving its intended purpose.

As we continue down the path of greater and greater federal power, the electoral college makes less and less sense.

(For what it's worth, I am in favor of restricting the federal government to its proper constitutional role rather than abolishing the electoral college. )

EPG writes:

Perhaps you should reconsider your theories in light of real-world experience.

Parties in countries with popular vote elections don't usually have regional support figures as large as 85% or 90%, and the few that do are normally just representing their region, not being nationally competitive.

Consider France or Mexico, which have popular-vote presidential elections, or countries with PR legislative elections. They normally have broad social coalitions behind the biggest two or three parties, because economic and social interests create more common ground than geography.

The USA would probably follow their example; like most other rich countries, it has a liberal left-wing party and a conservative right-wing party.

And, of course, the electoral college didn't stop the last US civil war.

G in Australia writes:

An amazing discussion given that most of the rest of the democratic world does not have an EC, and therefore there is much data about how things behave in it's absence, but nothing in the post, and only one comment.

But agree about "democracy as a means to an end" - the main end is that the government can respond to the people enough that they do not come to exercising their will through the shedding of blood.

Dave writes:

So we should hope that voting is a stupid proposition* for the majority of Americans in the name of less social conflict? That's the best solution for a democracy? Also although the outcome is not in doubt for most states, very few states have over 60% share of votes for one party or the other. I think these regional differences are overstated. A plan to benefit a minority of regional concerns is still a minority. As it is now, we've randomly given Iowa a disproportionate amount of political influence, which leads to some of those farmer benefits.

What about the effect on third party candidates? In 1992, Ross Perot won nearly 20% of the popular vote. That should have been very encouraging for third party candidates, but he won 0 electoral votes, which made it discouraging. The system encourages the status quo of our two awful political parties.

Also, we HAD a civil war with the electoral college system in place, so perhaps its claimed social conflict prevention benefits are overstated as well...

What are you going to defend next? Having two Senators per state? (sorry, this was snarky, but I consider these two issues to be related archaic relics from the past that impede the political process).

Forget the historical origins. If you were starting a country from scratch here, would you implement an electoral college? Is social conflict really the rationalization you would use to sell it?

* cost of time vs probability of swing the national election * marginal benefit; the right side of that equation is essentially 0 for the overwhelming majority of individual voters currently.

Dave writes:

Not to mention that this means that you favor the influence of arbitrary intranational geographical borders to influence a national election...

Ken B writes:

What a cool argument.

Another plus for the EC (with the winner take all rule, which is actually a separate thing) is that it forces candidates to attend different, conflicting, claims. We know from studies of decision making this is usually a good thing for better decisions in gneral. But in line with Garett's main thesis, this alos makes it harder to assemble jsut a simple majority based on some claim or issue. Issue X might attract 53% of voters, and they might care enough to impose stern actions on the 47%. But with the EC unless that 53% is spead evenly, and it never is, that won't suffice.
I think this imposes a certain moderation and "let's get along" effect on elections.

Eli writes:

In a popular election, Democrats would have incentive to take a more 'extremely Californian' position to capture more votes there, but would also have incentive to start appealing to liberal and Hispanic Texans. (The reverse is even more true for Republicans). What reason do we have to think the former incentive is greater than the later? What evidence is there that inter-regional conflicts are stronger than nationally recurring intra-regional conflicts?
The current system may leave out regional conflicts, but it may also leave out large interest groups that are too spread out to sway any given state. So it seems plausible that regionally disparate national divides along income, race, and occupation could yeild a "more purple" map than we have now.

Adam Baum writes:

I find it fascinating how few people, including the author understand the real genius and purpose of the electoral college-fraud mitigation.

Let's assume that there's a fairly tight race, like this one. If there was no electoral college, one side or the other-and let's be honest-the Dems work in electioneering like Michaelangelo worked in clay (don't get me wrong, the other side tries, but are miserable)-would simply need to swing a few places. Voting early and often in Cook County Illinois, Boston Mass, and have a few Black Panthers show up in paramilitary get ups in Philadelphia.

As it stands now, you need a tight state for these tactics to work-and then they can only swing that state-and you don't get to counter votes in another state, because the total number is fixed in advance. Under a popular vote, the prize for fraud is the entire election.

C.S. Rodet writes:

I like this argument; however, the combination of the electoral college and the capability of mass marketing also increase the ambiguity in presidential campaigns. We tend to hear more about "Change we can believe in" than useful information.

Ron writes:

Very poor argument (with zero supporting evidence).

Why is the danger of potential "regional conflict" worse than the evils of the current system, which include:(i) the over-emphasis on the particular concerns of the individual swing states; and (ii) the unfairness of a Texan's vote counting 1/3 as much as that of a person from Wyoming? Unfortunately, Garrett doesn't even try to answer this question.

Nor does he address the issue raised by several commenters, i.e., the history of intense regional conflict (including a bloody civil war) that has existed under the current system.

Sure, the popular vote system wouldn't be perfect either but at least we'd be starting from a position of one person-one vote.

vidyohs writes:

Personally I view the "winner take all" way of administering the electoral votes as being one of the most gross violations of the electoral college. The "winner take all" interpretation destroys district representation in so many cases and also destroys the reflection of the popular vote.

If electors voted for the candidate their district chose, the election would reflect the popular vote as closely as possible. The elector for my district should reflect the choice my districts gives him, and if it goes against the majority of the other districts in Texas, so be it, and so what. To do otherwise denies us choice.

AmericanHeretic writes:

We have been heading in that direction for decades with the gradual vitiation of our Constitution, but when did America become a democracy?

L Nettles writes:

I recommend we experiment with Major League Baseball first. Let's change the world series to 7 games and the team that scores the most runs during the seven games wins the Championship.

Juan writes:

What bout senators and congressman? They are emphasizing regional their own.

Instead of geographic divisions, why not choose the first letters of the alphabet on someone's surname?

Jacob AG writes:

Three commenters have used the word "evidence so far." I would like to see some as well. What do we see happening in countries with a popular vote?

Airman Spry Shark writes:

I'm surprised no one's mentioned an obvious benefit of the EC: Sandy.

With a National Popular Vote, regional reduction in voter turnout (as Sandy is likely to do to the NE) would shift the influence over the election to unaffected regions for that election only (in this case, toward the red South & plains); the EC protects impacted regions' relative influence in the face of disruptive events.

I would assert that a nationwide implementation of the Maine-Nebraska system would enhance this (as well as the reduced incentive for fraud/intimidation); for instance, while the NE's aggregate influence is maintained in the face of Sandy, inland areas might be relatively disproportionate influence this time around.

Winner-take-all/first-past-the-post is still a problem, but not one the NPV addresses; that would require some kind of ranked/approval voting system (I prefer Condorcet methods).

tommy5d writes:

Civil wars usually occur between one group of median voters and another. You don't get a lot of long-lasting civil wars between 90% of the population and 10% of the population.

Eric writes:

"I recommend we experiment with Major League Baseball first. Let's change the world series to 7 games and the team that scores the most runs during the seven games wins the Championship."

No, I have a better idea. Let's leave the rules the same, but the winner of the World Series gets to pick the president!

matt writes:

This is a very interesting topic to me. I love reading the different opinions here. Might I suggest a different theory I am currently working on.
Rather than the EC going with a population/elector based system, change it up slightly, going with a state productivity based system. Electoral votes going to the states that produce neccessities. This way if you think about it, the states will constantly be changing to resources found rather than demographics. If you look at the production of goods for each state you will find the east and west coast provide very little in the way of energy, food and the things truly neccessary for the country as a whole. The mainly "republican sectors" or "red" are the "goods producing states as opposed to "service providers" in this great country. California is the only state that will exempt itself from this because not only is it high in population, it contributes to the energy and agriculture output that the rest of the country is needing for these population centers to heat and eat if you will, and allows them to vote. The sad part of the discussion is that the majority of people voting, vote with emotion and what benefits them personally as opposed to what will truly benefit their fellow man.
Altogether as a whole I really like this site.

Comments for this entry have been closed
Return to top