The Electoral College and the Case for Reform

The president is not chosen by the national vote total. The winner is whoever secures a majority of the Electoral College: 270 of 538 electors.

Each state's allocation is its House seats plus its two senators. The 23rd Amendment, ratified in 1961, gives Washington DC three — the number it would have as the least populous state.

Component	Number	Where it comes from
House-based electors	435	One per congressional district, reapportioned every ten years by census
Senate-based electors	100	Two per state, regardless of population
Washington DC	3	23rd Amendment (1961)
Total	538	Majority needed: 270

Two features do most of the work in the debates that follow.

Winner-take-all. In 48 states and DC, whoever finishes first — by one vote or by three million — takes every one of that state's electors. Maine and Nebraska are the exceptions, using the congressional district method: one elector for the winner of each district, two for the statewide winner. Both have split their votes in recent elections.

The two-senator floor. Every state gets two electors before a single one is allocated for population. This is a deliberate federal feature, and it is the root of the malapportionment argument.

The Constitution does not require states to hold a popular vote for president at all. Article II leaves the manner of appointing electors to each state legislature; the popular vote is a matter of state law and convention, not a federal requirement.

Five presidents have taken office having lost the national popular vote: 1824, 1876, 1888, 2000 and 2016. Two of those fall inside the specification's post-2000 window and are the examples to know.

2000 — Bush v Gore. George W. Bush won 271 electoral votes to Al Gore's 266, while Gore led the national popular vote by roughly 540,000. The outcome turned on Florida, where the certified margin was 537 votes out of nearly six million cast, and was settled when the Supreme Court halted the recount in Bush v Gore.

2016 — Trump v Clinton. Donald Trump won 304 electoral votes to Hillary Clinton's 227, while Clinton led the popular vote by roughly 2.9 million, or about two percentage points. Trump's College majority rested on narrow wins in Michigan, Pennsylvania and Wisconsin totalling under 80,000 votes.

The mechanism is the same in both cases: winner-take-all converts narrow state wins into whole blocs of electors, while surplus votes in a state the candidate was certain to carry are discarded. Clinton's four-million-vote margin in California added nothing to her College total beyond the 55 electors she would have won anyway.

Do not describe this as the system malfunctioning. It is the system operating exactly as designed — the presidency is a federal office elected by states, not a national office elected by individuals. Whether that design is still defensible is the argument the specification wants.

Because every state receives two electors regardless of population, an elector does not represent the same number of people everywhere.

Wyoming's 576,851 residents elect three electors — about 192,000 people per elector. California's 39,538,223 residents elect 54 — about 732,000 per elector. A Wyoming voter's share of the College is therefore close to four times a Californian's.

Defenders answer that this is the point, not a flaw: the Senate is malapportioned for the same reason, and the College inherits it. The United States is a federation, and small states accepted the constitutional settlement in 1787 partly because it protected them from being governed entirely by the large.

Arguments for keeping the Electoral College:

• It preserves federalism. The president is elected by the states, which matches a constitutional structure in which the states are genuine political units rather than administrative regions.
• It protects small and rural states from a system in which candidates could campaign only in dense population centres.
• It usually produces a clear, quick result. Narrow national margins are contained within individual states, so a recount is confined to one state rather than run nationwide — the 2000 recount was a Florida problem, not a fifty-state one.
• It rewards broad geographic coalitions, discouraging candidates who appeal intensely to one region alone.
• It reinforces the two-party system, since a third party that polls evenly across the country wins no electors at all.

Arguments for reform:

• It can install the loser of the popular vote, twice since 2000 — a legitimacy problem in a country that presents itself as a model democracy.
• Votes are not of equal weight, on the roughly four-to-one gap shown above.
• It concentrates the campaign on a handful of states. In 2024 the contest was fought overwhelmingly in Arizona, Georgia, Michigan, Nevada, North Carolina, Pennsylvania and Wisconsin. Voters in safe states are, in practice, spectators.
• Turnout suffers where the outcome is certain, because a vote for either party in a safe state cannot change the state's electors.
• Faithless electors introduce a small but real uncertainty. In 2016, seven electors voted for someone other than their pledged candidate — the most since 1872. In Chiafalo v Washington (2020) the Supreme Court held unanimously that states may enforce an elector's pledge, which narrows the problem without abolishing it.

Four proposals recur, and they differ sharply in how hard they are to achieve.

Reform	What it would do	What it needs
Direct national election	Abolish the College; the popular vote winner becomes president	A constitutional amendment
Proportional allocation	Split each state's electors in proportion to its vote	State law, or an amendment if imposed nationally
Congressional district method	Extend the Maine and Nebraska model to all states	State law, state by state
National Popular Vote Interstate Compact	Member states award their electors to the national popular vote winner	Member states reaching 270 electoral votes between them

The NPVIC is the only one making measurable progress. It is an agreement among states to pledge their electors to whoever wins the national vote, and it is written to take effect only once its members hold 270 electoral votes — at which point it decides the presidency without touching the Constitution. As of August 2026 it had been enacted by 18 states and the District of Columbia, holding 222 electoral votes, after Virginia joined in April 2026. It remains 48 short.

Its weaknesses are instructive. Every member so far is a reliably Democratic-voting jurisdiction, so it has yet to prove it can attract the competitive or Republican-leaning states it needs. It would also face immediate litigation: Article I gives Congress a role in approving interstate compacts, and a state that had pledged its electors against its own voters' choice would be sued at once.

The obstacle is Article V. A constitutional amendment requires a two-thirds majority in both chambers of Congress and ratification by three-quarters of the states — 38 of 50.

That threshold hands a veto to precisely the states with most to lose. The thirteen smallest states can block any amendment between them, and the two-elector floor is worth more to them than a national popular vote would be.

Partisanship compounds it. Because the College's bias has favoured Republicans in the two recent divergences, reform has become a broadly Democratic cause and resistance a broadly Republican one — so the two-thirds congressional majority is unreachable while the parties are closely matched.

The nearest miss came in 1969–70, when a direct-election amendment passed the House by a large majority and was then filibustered to death in the Senate. Nothing has come as close since.

The strongest exam judgement here is not "the College should be abolished" but an argument about feasibility versus desirability. You can hold that the College is indefensible in principle and still conclude that reform is realistically impossible, and that the NPVIC is the only route with any prospect of success.

The specification pairs the Electoral College debate with the role of incumbency, and the two are connected: both are about whether US elections are genuinely competitive.

Incumbents win. House incumbents seeking re-election are routinely returned at rates above 90%, and Senate incumbents usually above 80%. The advantages compound:

• Name recognition and a record to campaign on, built over a two- or six-year term
• Fundraising, since interest groups and PACs prefer to back the likely winner — an incumbent typically far outraises a challenger
• Casework and constituency service, paid for from official allowances
• Franking, the right to send official mail to constituents at public expense
• District boundaries, since state legislatures draw congressional districts and often draw them to protect sitting members

The consequence is a shortage of real contests. Most House seats are safe, so the meaningful election is the primary rather than the general — which pulls candidates towards their party's activist base and away from the centre, feeding the polarisation examined elsewhere in this topic.

Set against that: incumbency is not a guarantee. Mid-term elections routinely punish the president's party, and a wave election can remove dozens of incumbents at once. Congressional elections fall on 3 November 2026, with Republicans defending a narrow House majority and 22 of the 35 Senate seats contested.

Campaign finance and the influence of interest groups are the other two limbs of specification 5.4. They are covered in the campaign finance and interest groups lessons — in an essay on the health of US democracy, bring all four together rather than treating them separately.

1. Describing how the College works instead of debating it

Section 5.1.1 covers the mechanics of the process. Section 5.4 asks whether it is defensible. A 30-mark answer that explains how electors are allocated and stops there has answered the wrong question.

2. Saying the system "failed" in 2000 and 2016

It did what it was built to do. The argument is about whether a federal, state-based election of a national executive is still appropriate, not about a malfunction.

3. Treating abolition as a live prospect

It is not, and saying so is a stronger answer than pretending otherwise. Article V's 38-state threshold, plus the small-state and partisan interest in the status quo, is why the debate has moved to the NPVIC.

4. Getting the NPVIC threshold wrong

It does not need 38 states. It needs member states holding 270 electoral votes between them — a majority of the College, not a supermajority of states.

5. Asserting that the College always favours one party

It has favoured Republicans in the recent divergences, but the bias follows the geographic distribution of each party's vote and has shifted before. Date any claim you make about it.

6. Forgetting that incumbency is part of this section

5.4 lists incumbency alongside the electoral process. An answer on the competitiveness of US elections that omits the 90%-plus House re-election rate has left out required content.