WHEN AMERICA’S founders established the electoral college—the group of state-appointed electors who officially choose the president—the world was a very different place. The decision was partly a compromise over slavery, with southern states demanding that their slaves count as three-fifths of a person when apportioning votes, even though they could not cast ballots themselves. But it was also a product of contemporary aristocratic ideas about democracy and distrust of the masses.
America has evolved since 1787. Slavery has been abolished for 150 years. Most Americans cherish democracy and majority rule. The electoral college has evolved, too. The original model, in which the president is chosen by a “small number” of electors hand-picked to deliberate about the qualifications of the candidates, has been amended to allow states to allocate their electoral votes to the winner of the popular vote. Each state has one voter in the 535-strong college for every member it sends to the House of Representatives and the Senate.
Yet the electoral-college system, for all its reforms, may still fail to reflect the choice of most voters. In two of the past five presidential elections, the winning candidate has actually lost the popular vote (something which happened only three times previously, in 1824, 1876 and 1888). In 2016 Donald Trump attracted 2.9m fewer votes than Hillary Clinton and yet was awarded 77 more electoral votes. What matters is not only how many votes candidates get, but where they get them. Because electoral-college votes are assigned by state rather than by the country as a whole, and because smaller states receive a disproportionately high number of votes, it is possible for smaller states to overrule the majority of voters. We can quantify which party the system favours by calculating the difference in the winning candidate’s vote margin nationwide and in the “tipping-point” state that decides the election.
In 2016, that state was Wisconsin. There, Mr Trump’s vote share was 1.5 percentage points higher than his performance nationwide (excluding third parties). That was the highest gap since at least 1972.
A similar gap will probably reappear this year. The Economist’s new presidential election model puts Mr Biden on track to win 53.5% of the nationwide two-party vote, and 52% in Pennsylvania, the state most likely to determine the outcome in November. That yields a gap of 1.5 percentage points, which suggests that there is a significant chance that Donald Trump will repeat his electoral victory of 2016.
In fact, that is his most likely path to victory. Each day, our model generates 20,000 different scenarios for how the election might unfold. Each one of these “simulations” makes predictions for how the states will vote and tallies each candidate’s electoral-college and national popular votes. At the moment, Mr Trump wins the election in 15% of those scenarios. But he is nearly three times more likely to win the majority of electoral-college votes with a minority of the popular vote (an event with an 11% chance of occurring) than he is to win the majority of both (4%).