The Electoral College is a strange, patched-together constitutional contraption. But it poses a smaller long-term threat to American democracy than the Senate, because the problems is causes are less serious and there are plausible ways to address them. In contrast, the Senate undermines principles of equal democratic representation and we have no viable way to address most of these problems within our constitutional framework.
On the surface, these institutions seem to have similarities. They are both state-based institutions, reflecting the larger political importance of the states, less mobile population, and lower level of interstate economic integration in the 1780s. They both also allow states to use their political power on a winner-take-all basis, unlike the House of Representatives.
As the Electoral College operates now — except in Maine and Nebraska where some electoral votes are allocated by House district — the candidate who wins the plurality of a state’s popular votes receives all of its electoral votes. A state’s Senate representation is determined by two winner-take-all elections, now by popular vote, but before 1913 by state legislatures.
But that’s where the similarities end. They do not advantage the same states. They are not equally serious threats to democratic representation in the future. And they are not equally challenging to reform.
The first thing to know about the Electoral College is that, as originally designed, it was a disaster. It very quickly stopped working at all. The Electoral College as written in 1787 assumed presidential and vice presidential candidates would not run together on party tickets. Under the original rules, there is only one Electoral College vote, in which each elector gets two ballots, and the second-place winner becomes vice president.
In America’s first contested election in 1796, to succeed George Washington, the problems became apparent. Long story short, the Democratic-Republican Party nominated Thomas Jefferson for president and Aaron Burr for vice president and the Federalist Party nominated John Adams for president and Thomas Pinckney for vice president. In this system, if all the winning party’s electors voted for both candidates on its ticket, those candidates would tie, sending the decision to the House of Representatives.
In 1796, a substantial number of Jefferson’s and Adams’s electors voted for various other candidates with their second ballot. As a result, Jefferson got the second-most votes behind Adams. Neither running mate was elected to office.
In 1800, Jefferson ran with Burr again, opposing Adams’s bid for reelection. This time, Jefferson’s Democratic-Republican Party had more supporters in the Electoral College, but the Democratic-Republican electors all voted for Burr with their second ballot, throwing the decision to the House of Representatives. The House, which, as specified in the Constitution, votes by state, fell one state short of giving Jefferson a majority, with the other states voting for Burr or casting blank ballots because their representatives were evenly divided. Only on the 36th ballot did several Federalist representatives in Maryland and Vermont change their votes, throwing their states to Jefferson, leading to his election.
This system simply did not work. The original Electoral College wasn’t just a poor method of selecting a president and vice president. It was considerably worse than that. It couldn’t make any selection at all.
The system that we have now is not what was planned by the authors of the 1787 Constitution. It is that failed system after being patched up by the 12th Amendment, which in 1804 established separate ballots for president and vice president, and also after the winner-take-all method of assigning a state’s electoral votes spread to almost all states by the mid-1800s after initially being used by only Pennsylvania and Maryland.
For most of American history, the Electoral College has rarely mattered. Presidential candidates are more likely to campaign in, and cater to, close states (more on that later), but most of the time this strategizing ends up being unnecessary because the Electoral College selects the same winner as the popular vote would have. The exceptions are 1824 (where third parties threw the election to the House of Representatives), 1877 (where President Hayes might have won a majority of the popular vote if not for rampant suppression of the black vote in the former Confederacy), 1888, 2000, and 2016. Obviously, these last two examples loom large in our thinking.
The Senate has also gone through various iterations. Its members were chosen for six-year terms by state legislatures under the 1787 Constitution and, since the 17th Amendment took effect in 1913, by direct election. However, the way Senators vote in the chamber has evolved over the years not according to any coherent plan.
The biggest aberration is the filibuster. The authors of the original Constitution intended both Houses of Congress to vote by majority rule. James Madison specifically mentions this in Federalist #22 and #52. A previous question motion was removed from the Senate’s rules in 1806 in a move to clean up unnecessary and unused portions of the rules.
Despite the Senate’s rule for forcing an end to debate being removed in 1806 in a move to clean up various unused portions of its rules, the chamber still proceeded on a majoritarian basis in the 1800s, with minorities occasionally able to delay things but not permanently block bills. In 1917, the Senate adopted a “cloture rule” under which two-thirds of senators present could end debate. Only a few bills were blocked by a minority of Senators in the mid-decades of the twentieth century, although those were mostly crucial civil rights and anti-lynching legislation.
In 1975, the cloture threshold was lowered to three-fifths of all senators. In the decades following this, the prevalence of bills and nominations blocked by filibusters greatly increased. At the same time, Congress has created several exceptions to the filibuster rules, most notably the increasingly used “reconciliation” rules, which allow bills affecting the budget to be passed by majority vote, and, in the past decade, the abolition of supermajority requirements for all presidential nominations.
Because of their similarities — they are both unusual, state-based, winner-take-all constitutional features — it is easy to assume that the Senate and Electoral College both distort democratic representation in similar ways. But this is not the case. The Senate gives a big advantage to voters in small states, because every state gets an equal number of Senators.
Thus, California’s 39 million people get two senators in Washington, while two Senators also represent states like Wyoming (578,000 people), Vermont (626,000 people), and Alaska (737,000 people). In 2013, the New York Times pointed out that the six senators from California, Texas, and New York represented the same number of people as the 62 senators from the smallest 31 states. (Florida has since passed New York to be the third-biggest state, but the pattern persists.)
People in overrepresented states are not the same as the people in underrepresented states. While there are a few small states on the coasts (hello, Rhode Island and Delaware!), many more small states are inland and rural. The coasts and their large cities tend to be in larger states. This means that the economic and infrastructure needs of cities get less representation in the Senate.
America’s nonwhite population tends to be overwhelmingly in large or medium-sized states. To illustrate, the 10 biggest states (by 2018 Census estimates) all have nontrivial percentages of nonwhite voters, while the 10 smallest states mostly consist of rural, overwhelmingly white states.
While the Electoral College is also winner-take-all at the state level, each state’s representation is much more proportional to population. State are given electoral votes equal to their number of House members plus two senators. House seats are allocated to the states proportional to population. Only the extra two votes contribute to disproportionality.
When states have lots of House seats, the two extra electoral votes don’t have much effect on their overall share. But for small states, the two extra votes do give some boost, the biggest being in the states that have only three electoral votes (the minimum number), when in a strictly proportional allocation they would have less. Yet overall, this boost to small states over strict proportionality is much smaller than the complete equality small states get in the Senate.
Rather than small state voters, the Electoral College gives really disproportionate influence to voters on the winning side in close states, and less influence to voters on the losing side in close states and those in states where one party dominates. Evaluating which voters have an advantage in the Electoral College is similar to evaluating a gerrymander. In both, you are looking at the outcome of a series of winner-take-all “districts.” In a gerrymander, voters have more influence (and fewer wasted votes) if they win a lot of districts by a small amount, while their opponents are packed into a small number of districts where they win overwhelmingly. In the Electoral College, voters get more influence if they win states with a lot of Electoral Votes by small margins, while their opponents are packed into states where many votes are wasted because they win by huge margins.
The Republican Party won the Electoral College in 2000 and 2016, despite losing the popular vote. In these cases, the Republican candidates won narrow victories in states holding a lot of electoral votes. In 2000, George W. Bush won Florida, Ohio, Tennessee, New Hampshire, and Nevada all by less than four points — Florida famously by only a few hundred votes. In 2016, Donald Trump won Florida, Pennsylvania, Michigan, and Wisconsin all by less than two points, while Hillary Clinton ran up massive majorities in big states like California and New York (30- and 22-point wins, respectively). The Democrats’ Electoral College disadvantage was caused by narrowly losing states with a lot of electoral votes and running up large margins in states they won.
But these circumstances came about because of the very specific electoral patterns in these elections. They are not signs of a permanent disadvantage for coastal, urban, nonwhite voters, or liberal voters generally, in the Electoral College. There will probably be a lot of votes from these groups wasted in California for the foreseeable future. However, if Democrats could flip Florida, the Pennsylvania-Michigan-Wisconsin trio, or Arizona into to narrow wins, the efficiency of their vote distribution improves substantially. In 2012, for example, very narrow wins in Florida, Ohio, and Virginia led Barack Obama to win a much larger percentage of the electoral vote than the popular vote. And, of course, if there is a bit more of a regional realignment and Democrats ever narrowly win Texas, they could potentially have a big Electoral College advantage.
Because relatively small shifts in the location of the voting strength of the two parties can lead one party or the other to perform better in the Electoral College than the popular vote, it makes more sense to think of the Electoral College as introducing unpredictable random changes to election outcomes, rather than consistently favoring certain types of voters. This is not a good system. I certainly wouldn’t advise any country designing their constitution to adopt a system like this. But, with the distribution of voters and the trends we have now, it is not likely to consistently advantage some types of voters over others in the future.
Both the Senate and the Electoral College are strange constitutional relics, whose problems would be hard to fix. But the difference is that, for the Electoral College, there is a viable plan for curing its pathologies. Currently, 15 states with 189 electoral votes have passed the National Popular Vote Compact (NPV), under which states set in state law a policy that they will give all their electoral votes to the national popular vote winner. The NPV’s text says it will go into effect if enough states join to constitute a majority of Electoral College votes. The Compact is only 81 electoral votes short of a majority.
This plan seems constitutionally and legally viable. States have the legal power to set the rules for allocating their electoral votes, which is why Maine and Nebraska now don’t use the winner-take-all methods and allocate their votes partially based on congressional districts. Article 1, Section 2 of the Constitution says, “No state shall without the consent of Congress … enter into any agreement or compact with another state.” The National Popular Vote Compact needs to be approved by a majority vote in Congress to be constitutional, but this is much easier than trying to pass a constitutional amendment.
Seth Masket worries that state legislatures would pull out of the NPV when their electoral votes have to be given to a candidate who received less than a majority of their state’s popular votes. For instance, would Colorado’s legislature allow its votes to go to Donald Trump if he won its popular vote in 2020, or would they vote to pull out of the NPV? It is always hard to make predictions like this. Yet I think it is very possible that states would allow the NPV to govern their electoral votes even in this circumstance.
The key is changing expectations. Most ordinary voters don’t think about the Electoral College. Hopefully, they will think about it less and less the longer the NPV is in effect. The NPV will be the legal status quo in the states. Once the national popular vote system becomes the new norm among elites and the mass public, the national popular vote winner will have much more legitimacy than the losing candidate among the mass public and elites. I think it is plausible to expect state legislators to leave that status quo alone.
Compared to this, the challenges to fixing or abolishing the Senate are much bigger. There is no plausible fix like the NPV without a constitutional amendment. Furthermore, Article 5 of the Constitution states that “no state, without its consent, shall be deprived of its equal suffrage in the Senate.” This appears to imply that even an ordinary constitutional amendment couldn’t change the way Senate seats are allocated or abolish the body.
One reading of this is that, rather than the three-fourths of states that must approve regular constitutional amendments, every state would need to approve a change to Senate seat allocation. Given these limitations, we are left to nibble at the edges of the main problem. Here is a list of what could be done to improve the Senate and what would be required to implement the reform.
First, you could abolish the filibuster super-majority requirement in the last realm where it still exists: regular legislation not eligible for reconciliation. This could be done by a majority vote of the Senate in the same way that the filibuster on presidential nominations was ended in recent years.
Second, you could reduce the bias in Senate representation toward rural and white voters by admitting the District of Columbia and Puerto Rico as states. Both could be admitted by majority votes in both houses of Congress.
(There is some constitutional challenge with admitting DC, in that the admitting legislation would have to allocate a very small portion of DC as the remaining seat of government, because Article 1, Section 8 of the Constitution says there is space set aside for the seat of government, although it doesn't set a minimum size. Legislation could leave just the footprint of the Capitol Building as the seat of government. The legislation would also have to state that the District that is referenced in the 23rd Amendment, which gives DC electoral votes, refers to what is now the new state, not to the new smaller seat of government. A more detailed discussion of how this might work is too long for this article. Needless to say, there could be some challenges.)
After these first two reforms, the level of difficulty jumps considerably.
Third, even though changing Senate representation requires the unanimous consent of the states, you could pass an ordinary constitutional amendment that leaves the allocation of Senators the same but strips the entire Senate of some (or most) of its authority. The amendment could say that responsibility for judicial confirmations would switch to the House and some bills would no longer require Senate approval. However, an Amendment like this would require approval of two-thirds of both houses of Congress (including the Senate itself) and three-fourths of the states.
The fourth and ultimate way to solve the problem of the Senate would be to abolish it altogether or turn it into a body that, while smaller than the House, also allocates seats according to state population size. But as mentioned above, this is impossible. According to Article 5, this would require the approval of every single state.
The Electoral College is a constitutional nuisance that created big problems in 2000 and 2016 but poses fewer problems in the long run, even as it is now. We have plausible ways to reform it out of existence. In contrast, the Senate is a massive democratic problem with no plausible solution within our constitutional framework.
The Senate’s representational biases make it harder to do many things, including continuing to reduce systematic unequal treatment of nonwhite people in American society and trying to mitigate climate change. The most plausible reforms — ending the filibuster and admitting DC and Puerto Rico — only begin to reduce the problem. Anyone working to improve American public policy needs to think hard about the vexing problem of Senate reform, because without such reform, adequately addressing the most serious problems facing the United States is impossible.