National Journal

CVD homepage
What's new?
Online library
Order materials
Get involved!
Links
About CVD

A New Fix for the Electoral College
By Alan B. Morrison
January 3, 2003 

The 2000 presidential election focused the public’s attention on a number of serious flaws in our election system that have gone uncorrected for a long time. Congress has now passed legislation to deal with outdated voting machines and other practices that prevented tens of thousands of voters from having their ballots counted, but there remain three major problems that should be fixed.

First, every state except Maine and Nebraska employs the winner-take-all method of allocating electoral votes, which produced the situation in Florida where a few hundred votes out of more than 6 million determined who won Florida’s 25 electoral votes and became president. Second, in almost every presidential election since 1980, there have been third party candidates who received sufficient popular votes that, had they gone exclusively or even largely to the eventual loser, could have changed the outcome. The 2000 election made the role of third party candidates more visible, but there is no reason to believe that third parties will stop having a substantial impact on presidential elections. And third, the 2000 election could have wound up in the House of Representatives, where each state would have had a single vote, even though California had 52 House members and 8 states had only one. However one defines democracy, there is simply no defense in the 21st century to picking a president on a “one-state, one-vote” basis.

It would be convenient if Congress could pass a law solving these problems, but the Constitution makes that impossible. The first two issues are entirely within the control of the states, and unless all or a substantial majority reached a consensus on a solution, we would have confusion that would be as bad as the current situation. But even the states can’t fix the problem when no one wins a majority of the electoral vote because the Constitution mandates the current outcome. Since only a constitutional amendment can deal with all three, we should think big and decide what we want.

The most obvious answer is to do what every state does: elect the president by popular vote. If we were starting from scratch, the idea would have great appeal, but we are not. The Electoral College, despite its flaws, does reflect notions of federalism that are the basis of our system of government and would be hard to overturn. Because the Electoral College adds two votes for each senator to the one for each member of the House, the small states have more power than they would have under a strict population-based system. Since 24 states in the new Congress will have five or fewer House members, there is no chance that the requisite three-fourths of the states (38) would agree to abolish the Electoral College.

Historian Arthur Schlesinger, writing in The American Prospect [March 25, 2002, pp. 23-27] proposed a system under which the winner of each congressional district would receive one electoral vote, and the overall state winner would get two votes (corresponding to the number of senators). This is the choice that Maine and Nebraska have made, and the fact that no one else has joined them suggests it does not have strong support. Moreover, to make such a system work, it would have to be adopted nationally, and only a constitutional amendment could do that. If we are going that far, this kind of minimalist change seems hardly worth the effort.

The winner-take-all system should be changed not just because of what happened in Florida, but because it produces what is effectively a massive disenfranchisement of millions of voters who cast their ballot for a candidate who did not finish first in their state. For example, the more than 4.4 million Californians who voted for George W. Bush, constituting nearly 42 percent of the state’s electorate, might as well have stayed home since they produced not a single electoral vote for their candidate. The same can be said for the 8.3 million Bush supporters in Illinois, Michigan, New York, and Pennsylvania and the 4.5 million people who voted for Al Gore in Ohio and Texas. And in Florida, regardless of which candidate won, there would have been more than 2.9 million voters whose ballots might just as well have not been cast or counted. In these eight states alone, the winner-take-all effect negated 20 million out of a total of about 109 million votes cast nationwide. And, it seems highly likely that a significant, if undeterminable, number of voters in states where the race was not close simply did not go to the polls since their vote would have no impact.

Second, the winner-take-all rule forces candidates to focus on the undecided voter in the middle, and to do that, candidates often fudge their positions so as not to offend anyone and thereby lose their support. The blurring of distinctions among candidates, not to mention their efforts to hide or at least soften their true views until after the election, is bad for democracy. The more the candidates sound alike, the less inclined the mildly apathetic voters -- or those looking to shake up the status quo -- will be to go to the polls.

Finally, and most dramatically, the current system results in situations like Florida where a swing of fewer than 2,000 votes out of 6 million cast determined how all of Florida’s 25 electoral votes were awarded. Given human errors and the inevitable failures of even the most sophisticated equipment, having an election turn on so few votes should be avoided whenever possible.

The way to eliminate the possibility that a Florida problem will repeat itself, and to reinvigorate our presidential elections, is to institute a proportionate allocation system under which each candidate would receive the same percentage of a state’s electoral votes that he or she received in popular votes. Assuming a race between just Gore and Bush, Bush would have received 13 of Florida’s electoral votes, and Gore 12, meaning that any post-election fight would only be over one electoral vote, and would virtually eliminate the chance that any court, let alone the Supreme Court, would be in a position to decide the fate of the presidential election.

Changing from a winner-take-all system would also bring about significant changes in the way that candidates allocate their appearances and their expenditures, particularly for media. Once a state is seen as safe for one side or the other, no candidate spends any significant time or money there. But with proportionate voting, there is an incentive to campaign everywhere because, even in the states with the fewest electoral votes, it is relatively easy to pick up at least one of them. It is impossible to predict precisely how presidential races would change, but it is certain that they would not be the same, which hopefully would mean more interesting races and greater turnout.

What would have happened if a system of allocating electoral votes on a proportionate basis had been in place in 2000? There can be no definitive answer because it requires an assumption that both candidates would have conducted precisely the same campaigns as they actually did, and that would almost certainly not be the case. When the numbers were run using the actual votes cast, another problem surfaced that initially appeared to be a mathematical anomaly, but turned out to have far deeper implications. Because there were more than two candidates who received measurable amounts of votes – Bush and Gore each received more than 48 percent while Ralph Nader had about 2.69 percent and Patrick Buchanan had 0.44 percent -- under a proportionate allocation system, neither Bush nor Gore would have received the required 270-vote majority: Bush would have had 265 votes, Gore 263, and Nader 10. This raises the much more fundamental issue of what to do about third parties, but not just under a system of proportionate allocation of electoral votes, but under winner-take-all as well.

One choice would be to disregard the popular vote given to third parties and base the allocations solely on the votes of the two major party candidates, as, in effect, is done in a disguised manner under the current winner-take-all system. But that would not have resolved the problem in 2000 because the result would have been a flat tie at 269 electoral votes each, assuming no challenges or recounts.

Even if disregarding the 2.7 percent of the nationwide votes cast for Nader might have produced an actual winner in 2000, that approach would be a much more difficult sell in years like 1992 when Ross Perot had 19 percent of the popular vote and Bill Clinton and George Bush had 43 percent and 38 percent, respectively. Under the existing rules, Clinton garnered 370 electoral votes to Bush’s 168, but under a proportionate voting system, the totals would have been Clinton 236, Bush 197, and Perot 105-- again no winner. Since Perot actually finished second in Maine and Utah, and there were six other states where he won more than 25 percent of the vote, eliminating the votes cast for him would not be a legitimate option under a proportionate allocation system.

There are several ways to resolve a presidential election in which there is a significant level of support for a third party, but the fairest one, which would be most reflective of the will of the voters, is called “preference” or “instant run-off” voting and has recently been adopted by the City of San Francisco for choosing its mayor, council members, and other high-ranking officers. Each voter marks his or her first choice and then indicates a second or backup choice – but only if the voter chooses to do so -- which is counted only if no candidate secures an absolute majority of the 538 electoral votes (using proportionate voting). Had this two-prong system been in use in 2000, Nader and Buchanan voters who made a backup selection would have added their votes to those whose first choice was Bush or Gore to determine the eventual winner. In this way, voters would have been able to express their true preferences without fear that their vote might help elect the person they most strongly opposed. It might also have meant that, at least on the second round, the thousands of Floridians who mistakenly cast two votes for president might have had their votes counted, as long as at least one of their selections was Gore or Bush.

There is another important by-product of this inquiry: it requires us to think more deeply about the role of third parties in presidential and perhaps other elections. The instinctive reaction of the leaders of the two major parties (and probably many of their members) will be to oppose anything that gives third parties any degree of legitimacy. For them, proportionate voting and instant runoffs will only encourage third parties and thereby drain the strength and some of the money from their own party.

The real basis of that opposition is the assumption is that if the present system is retained, third parties will never again appear. But it is the current system that has spawned them, even with all the barriers that the major parties have placed in their way. Moreover, third parties are not a recent arrival, but have been a significant factor in presidential elections for well over a century, including the 1912 election when Theodore Roosevelt won as a third party candidate. It is not well known, but the Constitution neither enshrines the two-party system nor implies that presidential elections must be between only two candidates. Thus, both the original version of Article II, Section 1, clause three, and the changes to it made by the Twelfth Amendment adopted in 1804, specifically contemplated that more than two people will run for president. After requiring an absolute majority of the electoral vote to elect the president, the Constitution directs the House of Representatives to decide the election “from the persons having the highest numbers [of electoral votes] not exceeding three [originally five] on the list of those voted for as President.” Although the Constitution may not mandate greater accommodations for third parties in presidential races, it surely refutes the commonly held notion that the two-party system is an essential part of our democracy and that candidates who run against the major party nominees are somehow “un-American.”

But even if the major parties could continue the pretense that third parties don’t exist and therefore should be disregarded, they cannot hide from the fact that third parties do influence elections. Ralph Nader has been accused of being a spoiler who had no chance of winning, but took enough votes from Al Gore in Florida (and several other states) to give the election to George Bush. It is also true, though not widely known, that if Gore had won Florida, the same “spoiler” charge could have been leveled at Patrick Buchanan, even though he gained only 0.44 percent of the popular vote. If all of Buchanan’s supporters had gone for Bush in Iowa (7 electoral votes), New Mexico (5), Oregon (7), and Wisconsin (11), that would have added 29 electoral votes to Bush’s column, and Bush would have become president even if Gore had won Florida. Indeed, the spoiler effect is not limited to presidential elections. As John J. Miller of the National Review recently observed [New York Times, Nov 16, 2002, A27], the Libertarian Party has quite likely cost the Republican Party a seat in the Senate in each of the last three elections – this time in South Dakota and previously in Nevada and Washington.

Whatever truth there may to the claim that either Nader or Buchanan was a spoiler in the 2000 election, in some situations, there can be no doubt that a third party candidate, who cannot win, may draw enough votes from one of the other candidates to change the result, and that is true under the current regime or one based on proportionate allocation. But if that occurs, the blame should not be placed on the third party, but on the system that creates the potential for a spoiler effect. The proper question is not what we do about third parties, but what we do about the system so that third parties that will inevitably exist will play a constructive not a spoiler role. And the way to deal with that problem is not to let the initial vote be the final vote if no candidate obtains a majority, but to use instant runoff voting so that there will no longer be speculation about which way votes for a third party candidate would have gone.

Finally, it is time to change the way we decide presidential elections if no one garners an absolute majority of the electoral votes. Under the current system, if there is no winner, the election goes to the newly elected House of Representatives with each state having a single vote, regardless of the size of its House delegation. This means that it is at least theoretically possible that the 26 states with eight or fewer electoral votes could elect the president, even though they have only 28 percent of the electoral votes and their residents cast less than 18 percent of the popular vote in 2000. If there were instant run-offs, there would probably be no need for a backup system, but until such a change is put in place, we should change the deadlock system to eliminate the disproportionate role for small states, for example, by giving each member of the House, instead of each State, one vote.

Our system for electing a president may not be completely broken, but it is in serious need of repair. The winner-take-all outcome, the failure to account for the inevitable third parties, and a thoroughly outmoded way of finally resolving disputed elections should be changed. Let the debate begin.

(Alan Morrison is currently the Irvine Visiting Fellow at Stanford Law School. In late January, he will return to the Public Citizen Litigation Group, which he founded with Ralph Nader in 1972.)

Reform Options

Electoral College Table of Contents


top of page


______________________________________________________________________
The Center for Voting and Democracy
6930 Carroll Ave,  Suite 610, Takoma Park, MD 20912
(301) 270-4616        [email protected]