“The Real Preference of the Voters”

James Madison wrote a letter in 1823, the significance of which for electoral reform has been previously overlooked.

America is often described as a “Madisonian democracy” because we inherited our political and constitutional system from James Madison as its principal architect. While this Madisonian system has changed substantially over time—most importantly, the right to vote has expanded to encompass essentially all adult citizens—its core structural features, like federalism and separation of powers, remain its defining nature and distinguish it from other forms of democracy, like Britain’s parliamentary system.

Precisely because our democracy is Madisonian and not parliamentary, the choices we face when designing or reforming our electoral institutions are different from the options available to other democracies. We must elect our president one way or another, a task not faced in a parliamentary system with a prime minister chosen by parliament. Likewise, we must elect the Senate subject to the constraint that each state gets the same number of senators regardless of its population. As noted elsewhere, neither the president nor a state’s senator can be elected on the basis of proportional representation (PR), unlike a legislative chamber for which seats are allocated according to the principle of equal population per representative (what is often called “one person, one vote”). Thus, some other form of fair representative system is necessary for presidential and Senate elections, and given the preeminent role of the president and Senate in America’s democracy, the methods of electing these offices are especially important.

In deciding how to elect presidents and senators, we obviously are not wedded to Madison’s own views. Towards the end of his life, Madison himself wanted to change the method for electing presidents, even though he and his fellow allies of Thomas Jefferson had already altered the Electoral College system when adopting the Twelfth Amendment to the Constitution after tumultuous election of 1800. In Madison’s time, senators were elected by state legislatures rather than the people of each state. The Seventeenth Amendment changed that, and since 1913 citizens instead of politicians have chosen senators. It is up to us to conduct elections for presidents and senators in any way we choose that is compatible with the single-member nature of these offices (meaning that we fill these offices with one individual at a time and cannot simultaneously elect multiple individuals to fill multiple seats, as can be done in other electoral contexts).

Still, because our system is Madisonian, it is worth bearing in mind what Madison thought about how best to conduct elections. We have long understood that Madison generally favored majority-winner rather than plurality-winner elections, based on the elementary proposition that a majority’s preference was more likely to correspond to the interests of the public as a whole than the preference of a faction that numbered less than a majority. It was not believed, however, that Madison embraced the specific conception of majority-winner elections advanced by his French contemporary, the Marquis de Condorcet, who advocated that the proper method mathematically of determining the majority’s preference is to have voters reveal for each pair of candidates which one of the two they prefer.

It turns out that, at least late in life, Madison had thoroughly incorporated Condorcet’s analysis of elections into his own thinking on the subject. In 1823, Madison wrote a letter that reflects his understanding of Condorcet’s ideas. The significance of this letter and its invocation of Condorcet’s concepts has previously been overlooked, in part because Madison’s letter doesn’t mention Condorcet by name and because the relevant passage is short and an aside to the main point of the letter.

The influence of Condorcet on Madison is unmistakable, and I have written a paper to be published in the Wisconsin Law Review that explains the letter’s significance both for understanding Madison’s own political philosophy, at least as it evolved in his later years, and for our own efforts to carry forward our Madisonian system into the future. In the letter, Madison writes that a runoff with “the three instead of the two” candidates with the most votes (the emphasis is Madison’s) would have a better chance of electing the candidate who is “the real preference of the Voters” because “it not unfrequently happens, that the Candidate third on the list of votes, would in a question with either of the two first, outvote him.” This is Condorcet’s main point exactly: the candidate who finishes third in terms of being the first choice of voters is often the choice of a majority when compared one-on-one against each of the top two candidates.

After making this point in the letter, Madison dropped it for a reason that’s also derived from Condorcet’s work: the possibility that in an election with three candidates, none is preferred by a majority of voters when compared one-on-one to each of the other two. Madison believed that this possibility means that a three-candidate runoff is impractical. But subsequent scholarship building on Condorcet’s seminal work has developed practical tiebreakers for this kind of “Condorcet cycle.” In previous Common Ground Democracy essays, I’ve drawn upon that subsequent scholarship to show how it would be feasible to implement a Top Three system for both presidential and Senate elections. Implicit in Madison’s letter is that, if he had been able to know of this subsequent scholarship, he too would advocate for the adoption of a Top Three system to improve the likelihood of electing the candidate who, in his words, is “the real preference of the Voters.”

In the Wisconsin Law Review paper I explain why it is important, faced with the current crisis afflicting our Madisonian democracy, that we incorporate into it this kind of Top Three electoral system. The overriding aim of Madisonian constitutional architecture, with its separation of powers, is to protect the Republic from an authoritarian demagogue. Yet, as the paper details, our Madisonian democracy egregiously failed in this crucial respect by, first, electing Trump to the presidency in 2016 even though he was not then “the real preference of the Voters,” and, second, by failing to convict him in the trial of his impeachment for attempting to subvert the outcome of the 2020 election, which he lost, and, third, by electing him president in 2024 even though once again he was not “the real preference of the Voters.” (As discussed elsewhere, among the top three candidates in the 2024 presidential election, Nikki Haley was preferred by a majority of all the nation’s voters compared to either Trump or Kamala Harris one-on-one and thus, in Madison’s words, was “the real preference of the Voters.”)

The existing electoral system replicates the problem for Senate elections, also preventing in that context the candidate who is “the real preference of the Voters” from prevailing. As a result, the Senate has become much more beholden to Trump and thus fails to exercise its Madisonian responsibility of acting as a check and balance to the president. We are witnessing this failure this week, as senators like Joni Ernst are unable to withstand electoral pressure when voting on confirmation of Trump’s dangerous Cabinet nominees.

If we want to rescue our Madisonian democracy from its acute peril, we would do well to return to Madison himself and read his 1823 letter that invokes Condorcet’s electoral concepts and then figure out how best to implement them in a way that protects us from despots and their allies in office who are not “the real preference of the Voters.”