Document Type


Publication Date



Most criticism of the system of presidential election focuses on the Electoral College, and most criticism of the Electoral College focuses narrowly on the shortcomings of the Electoral College itself. The objections are well known. The most basic is an objection of political principle. The Electoral College, on its face, deviates from the democratic principle of one-person-one-vote and gives the vote of a citizen in Wyoming approximately the same weight as 3.5 votes in California. The result is an unequal distribution of political power, both between citizens and among states. We can call this the 3.5:1 problem.

There are also pragmatic worries about things that could go wrong. There is the risk that the winner of the national popular vote will not be the winner in the Electoral College: the wrong winner problem. There is the risk that one or more of the human electors will, in a moment of independence, seize the opportunity to vote contrary to their pledge: the faithless elector problem. Forty-eight states award their electoral votes as a block, creating an additional democratic imbalance: the winner-take-all problem.

These shortcomings, and others like them, are the subject of a vast literature, and the Electoral College is, by a considerable distance, the clause of the Constitution that has generated the most proposals for amendment. Let me stipulate the obvious: the criticisms seem to me valid, and reform is desirable.

But the Electoral College may not be the most urgent problem. This symposium was held shortly after the events of January 6, 2021, which demonstrated the necessity of paying attention to other parts of the system of presidential election. The Electoral Count Act of 1887 has come in for scorching criticism, as has the administration of elections inside the states. Those are major problems but have received so much recent attention that I shall leave them to one side. I wish to call attention instead to a different but overlapping and equally severe group of problems. The Electoral College is embedded within a process that starts well before the Iowa caucuses and continues, if necessary, to what happens after the electoral votes are counted on January 6: that is, to the “contingent procedure” in the House of Representatives. Proposals for electoral reform often take a narrow focus, tacitly assuming that the College itself can be eliminated while leaving the rest of the system unchanged. That may be correct, but it cannot be taken for granted, and questions need to be asked about potential effects elsewhere, and especially on the outer extremities of the process. Alexander Bickel made that observation in a trenchant book he published fifty years ago; the issues he raised need further exploration.

I shall discuss three questions: (1) What is the relationship of the Electoral College to the early stages of the election campaign? (2) What would be the effect of abolishing the Electoral College on the two-party system? (3) What is the relationship of the two-party system to the very last step in the process, the contingent procedure in the House of Representatives? These are not, of course, the only questions that could be asked about the interactive effects of the Electoral College with other parts of the process, but they provide a useful point of entry.

The crucial link here is the two-party system. Indeed, it is important to emphasize that, among the world’s constitutional democracies, the United States is anomalous in two ways: it is the only one to employ anything resembling the Electoral College, and it is the only one to have a deeply entrenched two-party system. Are these two things connected? Historically, without a doubt. Bickel thought they were structurally connected as well: that if you lose the Electoral College, you also lose two-party democracy. I do not think his argument succeeds, but the problem is exceptionally difficult.

As I shall explain, the three questions are interdependent, and the answers are anything but clear. The effect of abolishing the Electoral College on the national presidential campaign—that is, the campaign after the nominating conventions—can perhaps be guessed at, but the effect on the party primaries is hard to estimate and could have profound consequences for the nature of the political parties themselves. As for the two-party system, its roots, both historical and theoretical, seem to me extremely poorly understood. The system has been around for so long that it is taken for granted as almost an immovable object. But nothing in the Constitution or in the laws of political science mandates two parties, and the entire construct is (I think) more unstable than is commonly supposed.

Whether, in general, two-party democracy is to be preferred to multiparty democracy is not a question I attempt to decide. Multiparty democracy works extremely well in some of the world’s most stable democracies. Perhaps it would work here, too. But my point is a different one. The consolidation of the two-party system is the principal reason the contingent congressional procedure has not been used since 1825 – a time when Jefferson, Adams, and Madison were still alive. Although there have been a couple of close shaves, the two-party system has ensured that one of the two candidates wins an outright majority in the Electoral College. Calls for reform of the congressional procedure have therefore been rare. Why worry about a problem that never occurs? But if the delicate balance were to be disrupted—if (as the Framers expected) presidential elections were routinely sent for resolution to Congress—the consequences would be cataclysmic. That is my principal point. January 6 came perilously close to sending us along that path. Reform of the Electoral College itself can perhaps be postponed. Reform of the contingent congressional procedure cannot.


Constitutional law, presidential elections, party politics, Electoral College, two-party democracy

Publication Title

University of Pennsylvania Journal of Constitutional Law

Publication Citation

24 U. Pa. J. Const. L. 967 (2022)