The latest enthusiasm from progressive pundits and activists for replacing the American system of self-government is to abolish the Electoral College and choose presidents by national popular vote. As with all such enthusiasms — expanding the Supreme Court, abolishing the filibuster and the Senate itself, lowering the voting age to 16, letting convicted felons and illegal aliens vote, adding D.C. and Puerto Rico as states, automatic voter registration, abolishing voter ID, etc. — the scarcely concealed argument is that changing the rules will help Democrats and progressives win more.
Also as with all such enthusiasms, Democratic presidential contenders have been unable to resist its siren song. Multiple prominent Democratic senators, including Kirsten Gillibrand (N.Y.), Minority Whip Dick Durbin (Ill.), and Dianne Feinstein (Calif.), the ranking Democrat on the Judiciary Committee, are introducing a proposal this week in the Senate to make it happen, the second such proposal by Senate Democrats this month. As radical an idea as this is, its support in high places demands to be taken seriously.
The Electoral College has been with us since the Founding, and in its present form since the election of 1804. Some of the reasons for its creation may be obsolete now, and the original concept of the electors themselves as important actors in the presidential selection process has long since left us. But the fundamental system of electing presidents by 50 simultaneous statewide elections (plus D.C.) rather than a raw national popular vote has long served America well. It isn’t going anywhere, and it shouldn’t.
Uniting the States of America
What would American politics look like without the Electoral College? Changing our current system would unsettle so many of the assumptions and incentives that drive presidential politics that the outcomes could easily be unpredictable. But first, consider the immediate changes.
The core function of the Electoral College is to require presidential candidates to appeal to the voters of a sufficient number of large and smaller states, rather than just try to run up big margins in a handful of the biggest states, cities, or regions. Critics ignore the important value served by having a president whose base of support is spread over a broad, diverse array of regions of the country (even a president as polarizing as Donald Trump won seven of the ten largest states and places as diverse as Florida, Pennsylvania, Arizona, West Virginia, Wisconsin, North Carolina, and Texas).
In a nation as wide and varied as ours, it would be destabilizing to have a president elected over the objections of most of the states. Our American system as a whole — both by design and by experience — demands the patient building of broad, diverse political coalitions over time to effect significant change. The presidency works together with the Senate and House to make that a necessity. The Senate, of course, is also a target of the Electoral College’s critics, but eliminating the equal suffrage of states requires the support of every single state. A president elected without regard to state support is more likely to face a dysfunctional level of opposition in the Senate.
Consider an illustrative example. Most of us, I think, would agree that 54 percent of the vote is a pretty good benchmark for a decisive election victory — not a landslide, but a no-questions-asked comfortable majority. That’s bigger than Donald Trump’s victory in Texas in 2016; Trump won 18 states with 54 percent or more of the vote in 2016, Hillary Clinton won 10 plus D.C., and the other 22 states were closer than that. Nationally, just 16 elections since 1824 have been won by a candidate who cleared 54 percent of the vote — the last was Ronald Reagan in 1984 — and all of them were regarded as decisive wins at the time.
Picture a two-candidate election with 2016’s turnout. The Republican wins 54 percent of the vote in 48 states, losing only California, New York, and D.C. That’s a landslide victory, right? But then imagine that the Republican nominee who managed this feat was so unpopular in California, New York, and D.C. that he or she loses all three by a 75 percent–to–25 percent margin. That 451–87 landslide in the Electoral College, built on eight-point wins in 48 states, would also be a popular-vote defeat, with 50.7 percent of the vote for the Democrat to 49.3 percent for the Republican. Out of a total of about 137 million votes, that’s a popular-vote margin of victory of 1.95 million votes for a candidate who was decisively rejected in 48 of the 50 states.
Who should win that election? This is not just a matter of coloring in a lot of empty red land on a map: each of these 48 states is an independent entity that has its own governor, legislature, laws, and courts, and sends two senators to Washington. The whole idea of a country called the United States is that those individual communities are supposed to matter.
This scenario is extreme, but the problem is not: elections where some places are overwhelmingly for or against a candidate while the rest of the country is competitive. The most extreme example happened in 1860: Abraham Lincoln won 18 of the 33 states (all the free states except New Jersey), giving him 59 percent of the Electoral College. Across the states he won, Lincoln got 54 percent of the popular vote, just as in our example above. In the two states that gave electoral votes to Stephen Douglas (New Jersey, which split its votes between Lincoln and Douglas, and Missouri), however, Lincoln won just 26 percent, and he was not even on the ballot elsewhere: He got just 0.9 percent in the eleven states carried by Vice President John Breckinridge and 0.7 percent in the three states carried by John Bell.
Lincoln won a popular plurality with just under 40 percent of the vote, and it is true that 1860 is a unique case. But the point is that the Electoral College works against a united regional minority, such as the antebellum South, that seeks to impose its will on the majority regions of the country simply by virtue of superior unity.
In the case of the South, that unity persisted long after the Civil War. Well into the 20th century, in elections still within living memory, states in the “Solid South” voted in far greater lockstep than elsewhere. Democrats won Mississippi, for example, with over 82 percent of the vote in every election from 1892 to 1944, clearing 90 percent eight times. Franklin D. Roosevelt won the eleven states of the old Confederacy with 81 percent of the vote in 1932 and 1936, 78 percent in 1940, and 72 percent in 1944. But 94 percent of the vote in Mississippi and 81 percent in Alabama counted no more than Tom Dewey’s winning 50–49 in Ohio and Wisconsin in 1944, or Wendell Willkie carrying Michigan by three-tenths of a point in 1940. That’s a good thing if you think a single, radicalized region of the country shouldn’t be given disproportionate power in choosing a national leader.
One of the reasons we are having this argument right now is that more than 13 percent of Hillary’s voters lived in a single state, California — the highest proportion for any candidate since Dewey in his home state (New York, then the nation’s most populous) in 1944, and higher than any winning candidate since 1868, when only 34 states voted (a few ex-Confederate states were still not allowed to participate). Hillary’s 4.2-million-vote margin in California more than accounted for her 2.9-million-vote plurality nationally. That one-party unity in the largest state, out of step with the rest of America, explains more about the popular/electoral vote split than the small states do. In the smallest states (those with 5 or fewer electoral votes, including D.C.), Trump got 30 electoral votes to Hillary’s 29. The real Democratic grievance is not that small states get a voice, but that big, closely divided states such as Florida, North Carolina, and Ohio get a say instead of being swamped by a few big outliers.
Making It Work
That’s just how the vote tallies matter. But of course, replacing the Electoral College would change how votes are courted and how they are tabulated.
The immediate question is what happens when no candidate gets a majority of the national popular vote (i.e., more than 50 percent). Senator Jeff Merkley (D., Ore.), the author of one proposal, ducks the issue, claiming incorrectly that “we have now seen two elections where the majority of voters supported a candidate who did not become the President.” The proposed constitutional amendment authored by Senator Brian Schatz (D., Hawaii) and joined by Gillibrand, Durbin, and Feinstein states: “The pair of candidates having the greatest number of votes for President and Vice President shall be elected.”
In 49 elections since 1824, the voters have returned a popular-vote majority 31 times. Those elections have not been the problem: In only one of them (1876) did an arguable winner of a majority fail to win the Electoral College. I say “arguable” because Democratic New York governor Samuel Tilden’s 1876 race against Republican Ohio governor Rutherford B. Hayes was marred by force and fraud throughout the South that left many of the vote tallies in question, and one state won by Hayes (Colorado) held no popular vote at all because it had been admitted to the Union that year and was not ready for a statewide election.
In the other 18 elections, an Electoral College majority went to the winner of a plurality of the national popular vote 14 times. Those winners, all below 50 percent of the popular vote, include the first election of some very big names in American history: Lincoln, John F. Kennedy, Harry Truman, James K. Polk, Richard Nixon. Three Democrats won the presidency twice without ever winning a popular majority: Bill Clinton, Woodrow Wilson, and Grover Cleveland. Several of those winners fell far short of a majority, usually when there was a strong third-party candidate: 39.7 percent for Lincoln in 1860, 41.8 percent for Wilson in 1912, 43.4 percent for Nixon in 1968, 43.0 percent for Clinton in 1992. Yet the system legitimized them as the victors.
The Constitution currently provides that if no candidate wins a majority of the Electoral College, the president is chosen by the House of Representatives, with one vote per state. The one time that happened under the current voting system, in 1824, was a fiasco: the popular-vote winner (though in some states there was no popular vote), Andrew Jackson, was shut out when the third-place finisher, House speaker Henry Clay, threw his support to John Quincy Adams and was rewarded by becoming Adams’s secretary of state (then considered a stepping-stone to the presidency). That sort of deal is completely routine in parliamentary systems, but the American people rebelled and in 1828 chose Jackson by a clear majority after four years of a hobbled presidency. If the House were to choose presidents by the same method when there is no popular majority, George W. Bush and Donald Trump would still have won their elections, since the Republicans controlled of a majority of House caucuses at the time of each election. But the Democrats are proposing a system in which the president needs to win neither a popular majority nor a majority of the states. The Schatz amendment’s only fallback provision is that Congress can provide for resolving a popular vote tie.
If we think of the Electoral College as a way of ensuring a decisive result in the absence of a national popular majority, and election by the House as the fallback only when both of those options fail, it makes a lot more sense. Splits between the popular vote and Electoral College winners will practically always be very rare when one candidate gets a popular majority; 1876 aside, those splits have only happened when neither candidate mustered a majority. Hillary Clinton won 48 percent of the vote; had she won another 2.7 million votes away from Trump and the third-party candidates to get to 50 percent, the odds are good that she would have won the election. The combined margin of her defeats in the eight closest states — Michigan, Wisconsin, Pennsylvania, Florida, Arizona, Iowa, North Carolina, and Georgia — was just 813,700 votes.
Assuming that “national popular vote” means that even a plurality will do, what happens if the national popular vote is really close? We’ve had four elections (1880, 1884, 1960, and 1968) in which the popular-vote margin was less than a percentage point but the clear Electoral College outcome made it academic to reexamine the national vote totals. National Review’s editorial defending the Electoral College neatly summarizes the practical problems that would come swiftly to the fore:
Under the current system, the result of presidential elections tends to be clear almost immediately — there is no need, for example, to wait three weeks for California to process its ballots; it is nigh-on impossible for voters to return a tie or disputed outcome; and, because presidential elections are, in effect, fifty-one separate elections, accusations of voting fraud and abuse hold less purchase than they would if all franchisees were melted into a single, homogeneous blob. The freak occurrence that was Bush v. Gore is often raised as an objection against the status quo. Less attention is paid to the obvious question: What if that recount had been national?
All the various controversies over voting — who’s eligible, how do we check eligibility, how are votes counted and recounted, how quickly are they tabulated, what paper trail is created — have the advantage of being contained today within individual states, so that the states where election systems are dominated by one party can’t do more in a presidential race than deliver their own state. The diversity of state electoral systems and eligibility rules was a concern all the way back to the creation of the Electoral College in 1787, and overhauling the system would also create immense pressure to nationalize those systems, at the cost of much expense and controversy. That, too, is what progressives aim for: the Schatz amendment introduces a new rule that ties the right to vote in presidential elections to state requirements for voting for the state legislature, and federalizes in a stroke many aspects of election management: “The times, places, and manner of holding such elections and entitlement to inclusion on the ballot shall be determined by Congress.” There is no reason to think this would be worth the effort just to deal with the four cases in 231 years when a popular-vote plurality loser won an Electoral College majority.
Moreover, the Schatz amendment also grants de facto backdoor statehood to Puerto Rico and other territories for purposes of voting in presidential elections: “The President and Vice President shall be elected by the people of the several States, the territories, and the district constituting the seat of government of the United States.” This is entirely unprecedented, and would fundamentally alter the nature of territorial status as it has existed since 1787.
In fact, moving to a national popular vote could spell the end of popular-majority presidents. The 50-state system is a barrier to entry for third-party candidates with niche appeal, but a national popular-vote system could create quite different incentives. Michael Brendan Dougherty has elaborated on the destabilizing effect that Electoral College reform, especially combined with other proposals, could have on our existing two-party system. While some may view this as a good thing — everyone has their favorite grievances with the co-opting and polarizing two-party system — parliamentary systems with a profusion of parties tend to have a lot more parties that are openly extremist on the far right and far left. Allen Guelzo notes, more broadly, how the case against the Electoral College tends to ignore the experiences of other nations:
The German federal republic, for instance, is composed (like ours) of states that existed as independent entities long before their unification as a German nation, and whose histories as such have created an electoral system that makes our “antiquated” Electoral College look like a model of efficiency. In the German system, voters in 299 electoral districts each cast two votes in elections for the Bundestag (Germany’s parliament): the first for a directly elected member and the second for one of 34 approved parties (in 2017), whose caucuses then identify candidates. A federal president (Bundespräsident) is elected every five years by a federal convention that reflects the party majorities in the Bundestag and the state parliaments of the 16 German states. Finally, the federal president proposes the name of the de facto head of state, the chancellor (Bundeskanzler) to the Bundestag. By contrast, the Electoral College is remarkably straightforward.
As Jonah Goldberg observes, “no Western European country has a unified leader directly elected by the people.” Jonah continues:
As Elaine Kamarck of the Brookings Institution has observed, America is the only advanced democracy that has decided to strip its political parties of the power to select their own candidates. Until 1972 — through conventions, smoke-filled rooms, etc. — the parties, not the voters, determined who their presidential candidates would be. This function is among the informal checks and balances that make democracy workable around the globe; we scrapped it in favor of ever more open primaries.
The primary system in both parties today is still a state-by state system, and in the Democratic party it yielded a winner in 2008 (Barack Obama) who had not received the most votes nationally in the primaries, but went on to win a popular majority in the general election. It also answers a common argument that the Electoral College causes a lot of places to be ignored: The primary process still gives many of these same voters a big voice. Northeastern blue-state Republicans played a pivotal role in nominating Trump; rural red-state African-American southerners played a decisive role in nominating Obama.
Expand the House
Finally, there is a solution available (admittedly, one with broad-reaching effects of its own) that would reduce the incidence of popular-vote/Electoral College splits without requiring a constitutional amendment and without doing violence to two centuries of American practice or eliminating the role of the states. That solution is to expand the size of the House of Representatives.
The House began with 65 members, but the Founding Fathers expected that it would expand over time, with the addition of new states and the growth of population. The first census increased the total to 99, and the House grew steadily until it was frozen at 435 members in 1911. The Constitution puts a minimum size on House districts (30,000 people), but today the average district is more than 20 times that, exceeding 700,000 people per district.
That matters to the Electoral College because every state gets one electoral vote per House district, plus two for its two senators. In the early 1800s, it was common for only the newest state to have just three electoral votes, but today there are seven such states, plus D.C. Doubling the size of the House to 870 members would instantly increase all the states (but not D.C.) to a minimum of four electoral votes and would cut the “extra” electoral votes (those corresponding to Senators) almost in half, from 19 percent of the Electoral College to 10 percent. Based on the Census Bureau’s 2018 population estimates, doubling the House would, without any other changes, raise the four largest states from 28.4 percent to 30.6 percent of the Electoral College, while reducing the 15 smallest states and D.C. from 10.6 percent to 8.8 percent. That’s a fairer way of rebalancing the Electoral College without depriving the small states of a voice.
Expanding the House would affect many other things besides the Electoral College, of course: seniority, campaign finances, gerrymandering, even the physical capacity of the Capitol. So this is a remedy that shouldn’t be approached without serious thought for its other consequences. Moreover, the chief obstacle to doing so is that the people most opposed are the current members of the House. But if Democrats are serious about a realistic movement to improve the presidential election system, rather than just venting their anger at the fact that they have won a popular-vote majority only twice since 1976, they could use their new House majority to push for expanding the chamber in time for the drawing of new districts in 2022 and the presidential election of 2024.